Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Paste extrusion of polytetrafluoroethylene fine powder resins Ariawan, Alfonsius B. 2002

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2002-731227.pdf [ 6.04MB ]
Metadata
JSON: 831-1.0058968.json
JSON-LD: 831-1.0058968-ld.json
RDF/XML (Pretty): 831-1.0058968-rdf.xml
RDF/JSON: 831-1.0058968-rdf.json
Turtle: 831-1.0058968-turtle.txt
N-Triples: 831-1.0058968-rdf-ntriples.txt
Original Record: 831-1.0058968-source.json
Full Text
831-1.0058968-fulltext.txt
Citation
831-1.0058968.ris

Full Text

P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R R E S I N S by A L F O N S I U S B . A R I A W A N Master o f A p p l i e d Science (Chem. Eng.), T h e University o f Bri t ish C o l u m b i a , 1998 Bachelor o f A p p l i e d Science (Chem. Eng.), T h e University o f Bri t ish C o l u m b i a , 1996 A THESIS SUBMITTED IN PARTIAL F U L F I L L M E N T OF T H E REQUIREMENTS FOR T H E D E G R E E O F DOCTOR OF PHILOSOPHY in die Faculty o f Graduate Studies Department o f Chemical and Bio-Resource Engineer ing We accept this thesis as conforming to the required standard T H E U N I V E R S I T Y O F BRITISH C O L U M B I A October 2001 © 2001 Alfonsius B. Ariawan In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of Cwo^ CJJL g ^ M j t K ^ The University of British Columbia Vancouver, Canada Date fttc no DE-6 (2/88) A B S T R A C T D u e to its high melting point and melt viscosity, polytetrafluoroethylene ( P T F E ) is processed by a number o f unusual techniques. These include paste extrusion. In P T F E paste extrusion, a free-flowing fine powder resin, having a typical individual particle diameter o f 0.2 p m , is processed wi th the aid o f a lubricating l iquid to form an extrudate o f considerable strength. T h e process is carried out at near ambient temperatures and is usually fol lowed by sintering. A l t h o u g h P T F E paste extrusion has been commercialized, little is k n o w n o f the fundamental mechanisms underlying the process. In this work, the fundamental theoretical and experimental aspects o f P T F E paste extrusion were studied. F ive resins o f different molecular structure were tested. Experiments were conducted using Instron capillary rheometers, equipped wi th barrels o f different diameters and dies o f various design. Analyses were performed using a differential scanning calorimeter ( D S C ) , scanning electon microscope ( S E M ) and micro-Raman spectrometer. The tensile properties o f paste extrudates were determined using a universal Instron mechanical testing machine. In addition, visualization experiments were performed to determine the pattern o f P T F E paste flow during extrusion. Pr io r to extrusion, the P T F E powder-lubricant mixture (paste) is preformed to produce a compacted cylindrical billet. The preforming behavior o f P T F E pastes was studied i n this work, i n order to identify and determine the effects o f important processing variables. It was found that the m i n i m u m preforming pressure and the duration required to ensure uniform paste compact ion are dependent o n the resin molecular structure. A n empirical relationship was established to illustrate this. In the rheological study, the effects o f various operating parameters were investigated. T o quantitatively describe the flow behavior o f P T F E paste, a one-dimensional mathematical mode l was developed, based o n observations f rom flow visualization experiments. T h e mode l takes into account the elastic-plastic (strain hardening) and viscous nature o f the material i n its non-melt state. Finally, the mechanism o f P T F E paste flow, wh ich involves the formation o f fibrils, was determined using S E M and verified using D S C . T h e properties o f the extrudates were also analyzed, i n terms o f fibril quantity and quality (i.e. fibril orientation and continuity). A balance between fibril quantity and quality was found to be necessary to ensure acceptable product quality, as illustrated through the effects o f various operating variables o n the extrudate tensile strength. ii TABLE OF CONTENTS ABSTRACT ii LIST OF FIGURES vi LIST OF TABLES xi ACKNOWLEDGEMENTS xii 1 FUNDAMENTALS 1 1.1 Introduction 1 1.2 Chemical Properties of PTFE 1 1.3 Commercial Productions of PTFE: Polymerization Techniques 3 1.4 Processing and Applications of PTFE 5 2 PASTE EXTRUSION: GENERAL REVIEW 10 2.1 Introduction 10 2.2 Experimental Aspects of Paste Extrusion 10 2.3 Mathematical Modeling of Paste Flow 15 3 SCOPE OF WORK 18 3.1 Introduction 18 3.2 Thesis Objectives 18 3.3 Thesis Organization 19 4 EXPERIMENTAL WORK 21 4.1 Introduction 21 4.2 Experimental Equipment 21 4.2.1 Preforming and Extrusion Equipment 21 4.2.2 Other Apparatus 24 4.3 Materials 26 4.3.1 PTFE Fine Powder Resins 26 4.3.2 Lubricants. 26 4.4 Experimental Procedure 27 4.4.1 Paste Preparation 27 4.4.2 Preforming 28 4.4.3 Paste Extrusion and Visualization Experiments 29 4.4.4 Extrudate Analysis 29 5 GENERAL CHARACTERISTICS OF PTFE PASTE EXTRUSION 30 5.1 Introduction 30 5.2 Characteristics of Extrusion Pressure Curves 30 iii T A B L E O F C O N T E N T S iv 5.3 General Procedural Characteristics 35 5.4 Conclusions • 38 6 PREFORMING BEHAVIOR OF PTFE PASTES 39 6.1 Introduction 39 6.2 Effect of Preforming on Extrusion Pressure 39 6.3 Density Studies 41 6.3.1 Effect of Preforming Pressure 41 6.3.2 Effect of Preforming Duration 46 6.3.3 Effect of Lubricant Concentration 47 6.4 Lubricant Migration Studies 48 6.4.1 Effect of Preforming Pressure 48 6.4.2 Effect of Preforming Duration 50 6.4.3 Effect of Lubricant Concentration 51 6.5 Predicting the Effective Preform Length 52 6.6 Effect of Preforming from Both Ends 55 6.7 Conclusions 55 7 MECHANISM OF PTFE PASTE FLOW 57 7.1 Introduction 57 7.2 Morphological Development During PTFE Paste Extrusion 57 7.3 Mechanism for Fibrillation 58 7.4 Pattern of Flow Deformation 61 7.4.1 Visualization Experiments 61 7.4.2 Radial Flow Hypothesis 62 7.5 Conclusions 65 8 RHEOLOGY OF PTFE PASTES 66 8.1 Introduction 66 8.2 Theoretical Considerations: Development of 1-D Mathematical Model 66 8.2.1 Paste Flow through a Tapered Orifice Die ( L / D a = 0) 66 8.2.2 Paste Flow through a Tapered Die ( L / D a * 0) 74 8.3 Effects of Extrusion Conditions 76 8.4 Effects of Die Design 78 8.5 Model Predictions 83 8.6 Conclusions 87 9 PROPERTIES OF PTFE PASTE EXTRUDATES 88 9.1 Introduction 88 9.2 Quantitative Descriptions of Fibrillation: The Issue of Fibril Quantity and Quality 88 9.3 Effects of Die Design 93 9.3.1 Effect of Die Reduction Ratio 93 9.3.2 Effect of Die Entrance Angle 94 9.3.3 Effect of Die Aspect ( L / D J Ratio 96 9.4 Effect of Resin Molecular Structure 99 P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T I - I Y L E N E F I N E P O W D E R R E S I N S TABLE OF CONTENTS v 9.5 Effect of Lubricant Concentration 101 9.6 Conclusions 102 1 0 CONCLUSIONS 103 10.1 PTFE Paste Extrusion: Project Wrap-Up 103 10.2 Contributions to Knowledge 105 10.3 Recommendations for Future Work 107 N O M E N C L A T U R E 109 BIBLIOGRAPHY Ill PASTE EXTRUSION OF POLYTETRAI'XUOROETHYLENE FINE POWDER RESINS LIST OF FIGURES Fig. 1.1 Partial phase diagram o f polytetrafluoroethylene (Sperati, 1989) 4 Fig. 1.2 Schematic diagram o f a P T F E molecule, showing the helical arrangement o f the fluorine atoms around the carbon backbone (Blanchet, 1997) 4 Fig. 1.3 Crystalline structure o f P T F E : (a) band, (b) crystalline slices, wh ich may slip during deformation via shear i n the disordered regions wh ich separate them, (c) hexagonal array o f P T F E molecules wi th in crystalline slice (Blanchet, 1997) 7 Fig. 1.4 Schematic diagram o f P T F E paste extrusion for a wire coating process (DuPont , 1997) . 9 Fig. 4.1 Schematic diagram o f the preforming unit used i n this work. It is essentially the test frame o f an Instron capillary rheometer 22 Fig. 4.2 Schematic diagram o f Instron capillary rheometer. The unit was assembled under the test frame shown i n F ig . 4.1 for paste extrusion. A blank die was used i n place o f the capillary die during preforming 23 Fig. 4.3 (a) D i e assembly for the 25.4 m m diameter capillary rheometer, (b) visualization die design 23 Fig. 4.4 Coordinate axes for the Raman backscattering configuration 25 Fig. 5.1 Typica l transient extrusion pressure response from a ram extruder during P T F E paste extrusion 31 Fig. 5.2 Ex t rus ion pressure responses wi th preforms obtained using different procedures. It can be seen that there is little effect o n the magnitude o f the pressure peak 32 Fig. 5.3 Ex t rus ion pressure responses obtained by stopping the extrusion process twice. D u r i n g the first stop, the die was removed from its assembly and immediately reattached. D u r i n g the second stop, the conical part o f the partially extruded preform that conformed to the shape o f the die was cut to result i n a flat surface 33 Fig. 5.4 Ex t rus ion pressure responses obtained by stopping the extrusion process a number o f times, for various waiting periods (see figure). The results show the fading memory effect o f P T F E paste 34 vi LIST OF FIGURES vii Fig. 5.5 Ex t rus ion pressure responses obtained wi th longer waiting periods between re-extrusions. Experiments were conducted using the 25.4 m m diameter barrel. T h e results indicate that the fading memory effect is i n fact reversible 35 Fig. 5.6 Effects o f changing extrusion rate during a single extrusion run 36 Fig. 5.7 Effects o f continuous extrusion using individually prepared preforms 37 Fig. 6.1 Effect o f non-uniform preform densification o n the transient pressure profile dur ing extrusion (performing pressure o f 2 M P a ) 40 Fig. 6.2 Photographs showing parts o f a typical P T F E preform: (i) top and (ii) bo t tom. T h e top por t ion is the end at w h i c h pressure is applied 41 Fig. 6.3 Densi ty variation along preforms for the high molecular weight (open symbols) and l o w molecular weight (closed symbols) resins. T h e preforming duration was kept constant at 0.5 minute. E r r o r bars, indicating the standard deviation o f duplicate runs, are shown only for one o f the plots for clarity reason. Similar levels o f reproducibili ty apply to the other results 42 Fig. 6.4 S E M micrographs showing a typical surface o f a perform wi th magnification o f (i) 80x and (ii) 10,000x 43 Fig. 6.5 To ta l average preform density for the three resins, normalized wi th respect to the individual resin SSGs , as functions o f performing pressure 44 Fig. 6.6 B u l k (volume) compressibility o f unlubricated resins 3, 4, and 5 at r o o m temperature under the application o f a changing pressure applied by means o f a piston mov ing at constant speed 45 Fig. 6.7 Densi ty variation along preforms produced at 0.75 M P a wi th different preforming durations for the high molecular weight resin (resin 5) 46 Fig. 6.8 To ta l average preform density for the high molecular weight P T F E resin as a function o f preforming duration. Preforming pressure was kept constant at 0.75 M P a 47 Fig. 6.9 Relaxation o f pressure wi th time after compression o f unlubricated resins. 48 Fig. 6.10 Densi ty variation along preforms (high molecular weight resin) and total average preform density as functions o f initial lubricant concentration. Preforming was carried out at 0.75 M P a for 0.5 minute 49 PASTE EXTRUSION OF POLYTETRAFLUOROETHYLENE FINE POWDER RESINS LIST OF FIGURES viii Fig. 6.11 Lubricant concentration along preforms produced by applying different levels o f pressure for 0.5 minute 50 Fig. 6.12 Lubricant concentration along preforms produced at 0.75 M P a wi th different preforming durations (high molecular weight resin) 51 Fig. 6.13 Lubricant distribution along preforms having different initial concentration o f lubricant (high molecular weight resin). Preforming was carried out at 0.75 M P a for 0.5 minute 52 Fig. 6.14 Densi ty variation and lubricant distribution along preforms having different lengths 53 Fig. 6.15 Length o f preform o f uni form density as a function o f resin S S G and preforming pressure 54 Fig. 6.16 Length o f preform o f uni form density as a function o f initial lubricant concentration for the high molecular weight resin (resin 5) 55 Fig. 6.17 Effect o f preforming from both ends o n density and lubricant distribution (high molecular weight resin) 56 Fig. 7.1 S E M micrographs o f resin 2: (a) before processing (virgin), (b) after preforming, and (c) after extrusion. Similar morphologies are observed for other pastes 59 Fig. 7.2 Schematic diagram illustrating the proposed mechanism for fibrillation: (a) compacted resin particles enter the die conical zone, (b) resin particles are highly compressed and i n contact wi th one another i n the die conical zone, resulting i n the mechanical locking o f crystallites, (c) upon exiting the die, particles return to their original spherical shape, and entangled crystallites are unwound, creating fibrils that connect the particles 60 Fig. 7.3 S E M micrograph o f resin 2 after shearing i n a parallel plate rheometer 60 Fig. 7.4 Values o f the first heat o f melting obtained from D S C analysis for pastes before and after extrusion under various conditions 61 Fig. 7.5 Veloci ty profiles obtained from visualization experiments: (a) i n the barrel, (b) i n the die conical zones wi th Ot — 30°, (c) OL — 45°, and (d) a = 90° 62 Fig. 7.6 Schematic illustration o f the "Radial F l o w " hypothesis. T h e hypothesis assumes the existence o f a virtual surface o f radius ras measured from the die apex, on which all paste particles m o v i n g towards the apex have the same velocity 63 PASTE EXTRUSION OF POLYTETRAFLUOROETI-TYLENE FINE POWDER RESINS LIST OF FIGURES ix Fig. 7.7 F l o w patterns i n the die conical zones as calculated based o n "radial f l ow" hypothesis wi th Dh = 9.525 m m , £ > = 75.4 m m ' / s , and (a) a = 10°, (b) a = 30°, and (c) C£ = 45°. T i m e increments were set arbitrarily 64 Fig. 8.1 (a) V o l u m e element and (b) its dimensions i n the conical zone o f a tapered die according to "radial f l ow" hypothesis 67 Fig. 8.2 Force balance o n vo lume element: (a) forces acting o n volume element, (b) radial contribution f rom the four normal forces 68 Fig. 8.3 N o r m a l and frictional forces acting o n surface element (a) top view, (b) side v iew 69 Fig. 8.4 Force balance o n volume element i n the die capillary zone 75 Fig. 8.5 The effects o f temperature and extrusion speed o n the steady-state pressure o f paste extrusion 77 Fig. 8.6 The effect o f lubricant concentration o n the steady-state extrusion pressure for resin 3. Sol id lines are mode l predictions wi th fitted parameters listed i n Table 8.2 78 Fig. 8.7 The effect o f reduction ratio o n the steady-state extrusion pressure for resins 1—4. Sol id lines are mode l predictions using the fitted parameters listed i n Table 8.1 79 Fig. 8.8 T h e effect o f die L/Da ratio o n the steady-state extrusion pressure at different reduction ratios for (a) resin 1, (b) resin 2, (c) resin 3, and (d) resin 4. Sol id lines are mode l prediction using the fitted parameters listed i n Table 8.1 81 Fig. 8.9 T h e effect o f die entrance angle o n the steady state extrusion pressure: (a) at different extrusion rates for resin 3, (b) at 75.4 m m 3 / s for resins 1—4. Sol id lines are mode l predictions using the fitted parameters listed i n Table 8.1. A l s o shown i n (a) is the prediction using the B e n b o w -Bridgwater equation (1993) 82 Fig. 8.10 The relative magnitudes o f the strain hardening and viscous terms i n E q . (8.19) as functions o f (a) volumetric flow rate, (b) die entrance angle, and (c) die reduction ratio, for resin 4 wi th 18 wt .% I S O P A R G ® a t 35°C 86 Fig. 9.1 Correlations between the relative differences i n A H m l o f pastes before and after extrusion and the steady state extrusion pressure (under various extrusion conditions). T h e fact that the differences i n A H m l are always positive indicates that the resin crystallinity is consistently lower after extrusion. The lines are drawn to guide the eye 90 PASTE EXTRUSION OF POLn'ETRAFLUOROETI-IYLENE FINE POWDER RESINS LIST OF FIGURES X Fig. 9.2 Typ ica l Raman spectroscopy results for (a) unprocessed powder and (b) paste extrudate 91 Fig. 9.3 Development o f fibril orientation wi th extrusion pressure during transient extrusion experiment Raman intensity ratio is defined as I p a r a i i c | / Ip c r p e ndicuiar a t 1383 cm" 1 92 Fig. 9.4 T h e effect o f die reduction ratio o n extrudate tensile strength. A l s o shown are the steady state extrusion pressures corresponding to each experimental run. Lines are drawn to guide the eye 93 Fig. 9.5 T h e effect o f die entrance angle o n extrudate tensile strength. A l s o shown are the steady state extrusion pressures corresponding to each experimental run. L i n e is drawn to guide the eye 95 Fig. 9.6 T h e effect o f die entrance angle o n extrudate diameter. L ine is drawn to guide the eye 96 Fig. 9.7 T h e effect o f die L/Da ratio o n extrudate tensile strength and diameter. Lines are drawn to guide the eye 97 Fig. 9.8 Pictures o f extrudates (resin 5) obtained using dies o f (a) L/Da — 0 and (b) L/Da= 10 under the same experimental conditions. N o t e the visual difference i n the extrudate surfaces. T h e same effect was observed wi th other extrudates 98 Fig. 9.9 T h e effect o f resin melt creep viscosity (molecular weight) o n extrudate tensile strength. Lines are drawn to guide the eye. F i l led symbols represent copolymer series and unfilled symbols represent homopolymer series 99 Fig. 9.10 T h e difference i n the degree o f crystallinity between unprocessed and processes pastes having different molecular weights (melt creep viscosity). Recall that a high melt creep viscosity implies a high resin molecular weight 100 Fig. 9.11 The effect o f lubricant concentration o n extrudate tensile strength. L i n e is drawn to guide the eye 101 PASTE EXTRUSION OF POLYTETRAFLUOROE'lTIYLENE FINE POWDER RESINS L I S T O F T A B L E S Table 4.1 Properties o f P T F E fine powder resins studied i n this work. Relative magnitude o f the resin molecular weight can be inferred proportionally f rom the melt creep viscosity data 26 Table 4.2 Properties o f I S O P A R G® solvent ( E x x o n Corp . , 1994) 27 Table 8.1 Values o f material constants and coefficients o f friction for the different pastes 83 Table 8.2 Values o f material constants for resin 3 wi th different lubricant concentrations 87 xi ACKNOWLEDGEMENTS I wish to express m y sincere gratitude to m y supervisor, Prof. Sawas G . Hatzikir iakos, for his guidance and support during the course o f this study. I sincerely thank D r . Sina Ebnesajjad for his mentorship, his many useful suggestions during our discussions and most importandy, for his friendship during the course o f this work, and for insisting that I visited D u P o n t Fluoroproducts , Wi lming ton , D E , U S A for a training program (twice). T h e experiences that I have had at D u P o n t have been most fun, useful and rewarding, both for this w o r k and for me, personally. I am also thankful for the excellent hospitality that D r . Ebnesajjad has extended to me while I was i n Wi lming ton . Thanks also to M r . J o h n M c A d a m for the training that he has provided me at D u P o n t . H i s useful discussions o n the practical aspects o f paste extrusion are very much appreciated. John , thank you for suggesting me to check "the phase o f the m o o n " regularly, to explain any "weird" experimental findings. A t times, I found it consoling. Thanks to D u P o n t Fluoroproducts for the financial support and the supply o f the polymer samples. I am also appreciative o f the many testing conducted by D u P o n t technicians. M y colleagues and ex-colleagues f rom RheoLab at U B C have helped me i n various ways. I wish to thank Evguen i E . Rozenbaoum, D i v y a Chopra and M a n i s h Seth for their helpful discussions and exchange o f ideas. Finally, I w o u l d like to extend a special note o f thanks to my parents for their love and continuing support, and to my wife, Fujanarti, and my son, Anton ius , w h o have been a source o f strength and motivat ion for success. xii CHAPTER 1: FUNDAMENTALS 1.1 Introduction Polytetrafluoroethylene (PTFE) is a material of great commercial value. Its unique properties have earned it special recognition in the plastic industry. Since its discovery by Dr. Roy Plunkett in 1938 (Plunkett, 1941), it has "revolutionized the plastic industry and led to vigorous applications not otherwise possible" (Plunkett, 1987). In this chapter, the chemistry and properties of PTFE are examined, and the various polymerization techniques involved in its production are discussed. The various fabrication methods for PTFE, that are rather unconventional for a thermoplastic polymer, are also reviewed. Particular emphasis is placed on the process of paste extrusion, which bears the most relevance to the present work. 1.2 Chemical Properties of PTFE PTFE is a member of the fluoropolymer group, having a chemical formula of [-(CF2-CF^n-]. It is a perfluorinated, high molecular weight polymer, having a straight chain molecular configuration. PTFE is chemically inert, resistant to heat, and has excellent electrical insulation properties, as well as a low coefficient of friction over a wide temperature range (Gangal, 1989; Gangal, 1994; Ebnesajjad, 2000). These properties can be attributed to its molecular structure, in particular the C-F bond. The two types of bonds that made up a PTFE chain, the C-F bonds and C-C bonds, are extremely strong (Cottrell, 1958; Sheppard and Sharts, 1969), causing PTFE to have excellent mechanical strength and resistance to heat. The fluorine atoms can be envisioned to wrap around the backbone of the C-C chain, providing a protective shield to the chain from chemical attack. The size and electronic state of the fluorine atoms is exactly right, accounting for the chemical inertness and stability of PTFE (Gangal, 1994; Ebnesajjad, 2000). The protective shield also reduces surface energy, causing PTFE to have a low coefficient of friction and non-stick properties (Zisman, 1965; Gangal, 1994). PTFE, with its thermal and chemical stability, makes an excellent electrical insulator. It does not absorb water and its volume resistivity remains unchanged even after prolonged soaking. The dielectric constant of PTFE also remains constant over a very wide temperature range (Gangal, 1994). 1 CHAPTER 1 - FUNDAMENTALS 2 Since PTFE is insoluble in many common solvents, its molecular weight cannot be determined by conventional techniques. In practice, the number average molecular weight (MJ is usually estimated from the standard specific gravity (SSG) of the polymer. Higher SSG implies greater crystallinity and hence, lower molecular weight (Gangal, 1994; DuPont, 2001). Due to the linearity of PTFE molecules, the crystallinity of a virgin PTFE resin may be as high as 92-98% (Gangal, 1989). As a result, the SSG of PTFE is high for a polymer, typically ranging from 2.1 to 2.3. Following the standard procedure for measuring SSG (ASTM D4895), the number average molecular weight can be determined from M„=5.97xl0 5 log 1 0 0.157 2306-SSG (1.1) for resins with SSG > 2.18 (Doban et al, 1956; DuPont, 2001). The calculated molecular weights for PTFE with SSG < 2.18 are quite large (probably unrealistic), due to the asymptotic behavior of Eq. (1.1) in this range. M n has also been correlated to the second heat of melting, or the heat of recrystallization of PTFE ( A H ^ , obtained using differential scanning calorimetry (DSC) (Suwa et al., 1973; Gangal, 1994). It was found that M„ = 3.39 x 10 1 3 .AH m 2 5 1 6 . (1.2) for A H ^ ranging between 13.8 J/g to 32.7 J/g. Typically, M n is in the 106 to 107 range (Gangal, 1994). It should be noted, however, that Eqs. (1.1) and (1.2) are valid only for homopolymer PTFE resins. Comparison of PTFE molecular weight, regardless of whether or not the resins contain other comonomers, can be made by considering the resin melt creep viscosity instead. The melt creep viscosity, as detailed in US Patent 3,819,594 (Holmes eta/., 1974), is higher for a higher molecular weight PTFE resin (DuPont, 2001). The melting point of virgin PTFE is 342°C (Sperati, 1989), which is high for a thermoplastic polymer. DSC analysis indicates that the melting point of PTFE is irreversible (Gangal, 1994). A previously melted (sintered) PTFE will show a lower melting point of 327°C, which is the value that is often reported in the literature (i.e. second melting point). PASTE EXTRUSION OF POLYri?I"lMFLUOROETHYLENE FINE POWDER RESINS C H A P T E R 1 - F U N D A M E N T A L S 3 This implies that cooling does not re-crystallize the chains back to the original virgin configuration, making the resin less crystalline. During melting, a volume increase of 30% is typical (Sperati, 1989). The melt is stable, since even at 380°C, the melt viscosity is relatively high at approximately 10 GPa.s. Due to this high melting temperature and melt viscosity, it is not possible to fabricate PTFE resin using conventional polymer melt processes (Gangal, 1994; Ebnesajjad, 2000). Besides the melting point, PTFE has other transition temperatures, two of which are particularly important. These are shown in the partial phase diagram of PTFE in Fig. 1.1 (Sperati, 1989). Under ambient pressure conditions, the first transition occurs at 19°C. This is an especially important transition point due to its proximity to the ambient temperature. Fig. 1.2 shows a schematic diagram of a PTFE molecule, depicting the helical arrangement of the fluorine atoms around the carbon backbone (Blanchet, 1997). At around 19°C, a PTFE molecule undergoes a slight untwisting from a 180° twist per 13 CF 2 groups to a 180° twist per 15 CF 2 groups. The chain segments change from a perfect three-dimensional order to a less ordered one, and this results in a volume increase of about 1.3%. The second transition occurs at 30°C. Above this temperature, the number of CF 2 groups per 180° twist remains the same at 15. However, the extent of disorder of the rotational orientation of molecules about their long axis is increased. The total volume change as temperature is increased from below 19°C to above 30°C can be as high as 1.8% (Clark, 1962; Blanchet, 1997). This change in volume affects the density of the PTFE resin. 1.3 Commercia l Product ion of P T F E : Polymerization Techniques In the production of PTFE, tetrafluoroethylene monomer (TFE) is polymerized in a highly exothermic reaction. It is essential that the molecular weight of the resulting polymer is extremely high. Low molecular weight polymer will not have the strength needed in end use applications and will be of little commercial value. Polymerization of T F E is carried out in an aqueous medium involving either one of two procedures. The first procedure is known as suspension polymerization. In this method, little or no dispersing agent is used along with an initiator in a medium, which is subjected to vigorous agitation. This initially causes a stable dispersion of PTFE for a brief period of P A S T E E X T R U S I O N O F P O L Y F E T R A F L U O R O E T I T Y L E N E F I N E P O W D E R RESINS C H A P T E R 1 - F U N D A M E N T A L S 4 time. The PTFE particles later precipitate due to the lack of dispersing agent and the presence of agitation. The resulting dried polymer is 'stringy, irregular, and variable in shape' and is known as granular resin. The particle sizes of granular resin and its powder flow property can be varied, depending on the end product requirements, by size reduction (cutting), or by mixing different grades (Gangal, 1994). 8 r~i—i—i—i—|—i—i—i—i—|—i—i—i—i—|—i—i—i—i—|—i—i—i—i—|—i—i—i—r 0 20 40 60 80 100 120 Temperature (°C) Fig. 1.1 Partial phase diagram of polytetrafluoroethylene (Sperati, 1989). Fig. 1.2 Schematic diagram of a PTFE molecule, showing the helical arrangement of the fluorine atoms around the carbon backbone (Blanchet, 1997). P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 1 - F U N D A M E N T A L S 5 The second technique of polymerization is called aqueous emulsion polymerization. The procedure involves sufficient dispersing agent along with an emulsifying agent and an initiator. Gentle stirring is usually employed to ensure dispersion stability. It is important that the dispersion is not only stable, so that the PTFE particles do not coagulate prematurely, but that it is also unstable enough to allow subsequent coagulation to form a fine powder. The amount of emulsifying agent influences the rate of polymerization and particle shape. The resulting dispersion may be stabilized using a noninonic or anionic surfactant, and concentrated to 60-65 wt.% solid by electodecantation, evaporation, or thermal concentration. It may also be subjected to a mechanical separation instead. In the latter case, a fine powder resin is obtained. The resin is susceptible to mechanical damage and hence, during separation, shearing is carefully avoided (Gangal, 1994). Although the two procedures result in the same high molecular weight PTFE polymer, the products are distinctly different. The granular product can be molded in various forms. However, the resin obtained from aqueous dispersion polymerization cannot be molded, but has to be fabricated by dispersion coating, in the case of the concentrated dispersion, or by paste extrusion, in the case of fine powder resin (Blanchet, 1997; Ebnesajjad, 2000). 1.4 Processing and Applications of PTFE It is not possible to process PTFE resin by melt techniques, due to its high melt viscosity. Techniques involving cold pressing and sintering have to be employed instead. These techniques resemble those used in metallurgy, but are unconventional as far as thermoplastics processing is concerned. Granular PTFE resin can be fabricated into an end product by pressing the resin in a mold at room temperature to produce a preform of the desired shape. Preforming serves to reduce voids between the resin particles (Ebnesajjad, 2000), which would otherwise render the final product mechanically weak. The preform is then sintered at a temperature above the melt temperature of PTFE to allow coalescence of particles into a dense homogeneous structure. During sintering, the rate of temperature rise is important to allow the temperature distribution to be as uniform as possible. After sintering, the product is cooled at a specific cooling rate, which is also important in influencing the mechanical properties of the product. Fast cooling will lead to lower crystallinity and, consequently, higher tensile strength and P A S T E E X T R U S I O N O F P O L Y T E T I t A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 1 - F U N D A M E N T A L S 6 elongation at break, and better flex life for the product (Blanchet, 1997; Ebnesajjad, 2000). O n the other hand, a slow cool ing rate w i l l result i n a product wi th greater creep resistance, lower permeability to gases and solvents, and lower residual stress and distortion. Sintering and cool ing may be done under pressure or freely. Pressure cool ing may reduce the tendency o f the product to distort, although it is more costly (Blanchet, 1997). In the product ion o f continuous moldings such as pipes and rods, the method described above is not appropriate. Instead, a reciprocating ram extruder is used (Ebnesajjad, 2000). Screw extrusion, wh ich is normally employed i n the processing o f most other thermoplastics, is virtually impossible i n the case o f P T F E due to the high melt viscosity o f the resin and the tendency o f the unmelted powder to be compacted by the screw. In ram extrusion, the resin is sequentially charged into the reciprocating extruder to be preformed. The product is then pressed d o w n a heated tube, where melt ing and coalescence occur. Granular P T F E resin is commonly used to produce tubes, seals and piston rings, bearings, gaskets, and other basic shapes (DuPont , 1996; Blanchet, 1997; Ebnesajjad, 2000). Aqueous dispersions o f P T F E are commonly applied as coatings, or formed into films or fibers. W i t h particle diameters ranging from 0.1 to 0.3 prn and concentrations between 3 0 % to 60%, the dispersion may be applied o n surfaces by spraying, or by flow or dip coating methods (Gangal, 1994; Blanchet, 1997). D r y i n g is then carried out at a temperature below 100°C, before baking at temperatures between 216°C to 316°C to allow the removal o f the wetting agent. Subsequently, the coating is sintered at a temperature higher than 360°C. Generally, surfaces are coated wi th P T F E to provide protection against chemical attack, or to provide a non-stick property to the surfaces (Blanchet, 1997). P T F E f i lm may also be produced, by casting the dispersion o n a smooth surface. Subsequent drying, baking and sintering o f the fi lm is fol lowed by the stripping o f the f i lm f rom the surface. T o form P T F E fiber, an aqueous dispersion is forced through a spinneret and into a coagulating bath. T h e fiber is then heated, sintered and hot drawn to develop strength (Ebnesajjad, 2000). Aqueous dispersion finds a number o f applications i n fabric coating, i n the product ion o f porous fabric structures, such as Gore -Tex , and i n other antistick applications (Blanchet, 1997). The processing o f P T F E fine powder resin is o f particular interest i n this work. Fine powder resin is susceptible to mechanical damage above the transition temperature o f 19°C. P A S T E E X T R U S I O N O F P O L r i T i T R A F L U O R O E T I - I Y L E N E F I N E P O W D E R RESINS CHAPTER 1 - FUNDAMENTALS 7 Figure 1.3 shows the crystalline structure o f a P T F E powder (Blanchet, 1997). B e l o w 19°C, shearing w i l l cause crystal units to slide pass each other. A b o v e 19°C, however, molecules are packed more loosely and shearing w i l l cause the unwinding o f molecules, creating fibrils (Mazur, 1995; Ebnesajjad, 2000). It is important that fine powder resin does not undergo pre-mature fibrillation. T h e presence o f fibrils before processing w i l l result i n excessive extrusion pressure and mechanically defective products (Mazur, 1995; Ebnesajjad, 2000). Because o f this, the storage and transportation o f fine powder resin are carried out at temperatures below 19°C (DuPont , 1991). (a) (b) (c) Fig. 1.3 Crystalline structure of P T F E : (a) band, (b) crystalline slices, which may slip during deformation via shear in the disordered regions which separate them, (c) hexagonal array of P T F E molecules within crystalline slice (Blanchet, 1997). F ine powder resin is processed through paste extrusion. T h e resin is initially mixed wi th a lubricating l iquid o f l o w boi l ing point, for ease o f removal at a later stage o f the process. Ideally, a hydrocarbon wi th surface tension below 18 m N / m should be used to al low complete wetting o f the powder (Gangal, 1994). However , since this is not practical, liquids wi th fairly l ow surface tensions are used instead. T h e amount o f l iquid i n the mixture may vary from 16 wt .% to 25 wt .% (Ebnesajjad, 2000). Shear-free mix ing is performed by simply rotating the container, at a temperature lower than 19°C, for the reasons discussed above. T h e l iquid acts as a lubricating agent and as a cushion between particles so as to eliminate mechanical damage, wh ich may occur as particles slide past one another (Mazur, 1995). The powder-lubricant mixture, wh ich is called a paste, is al lowed to age at r o o m temperature for several hours before extrusion. This allows the paste to equilibrate at a PASTE EXTRUSION OF POLYTETRAFLUOROETHYLENE FINE POWDER RESINS CHAPTER 1 - F U N D A M E N T A L S 8 temperature higher than the mix ing temperature, wh ich results i n the lowering o f the lubricant surface tension and viscosity, and hence, better wetting o f the powder particles (Ebnesajjad, 2000). T h e next step o f the process involves the pressing o f paste to produce a preform. This serves to eliminate air voids that w i l l render the final product mechanically weak (Mazur, 1995). A preforming pressure o f 2 M P a is typically used. The preform is then extruded i n a ram extruder through a die o f specific shape, depending o n the product requirement. In a wire coating process, a special annular (crosshead) die is used, wh ich allows the introduction o f wire into the extrudate. The extrudate is then passed through a dryer to evaporate the lubricating l iquid before being sintered i n an oven. Electr ical spark tests may then follow. In the product ion o f tubes and pressure hoses, a similar annular die is used. T o make tapes, the preform is extruded through a simple conical entry die and calendered to a flat, continuous product before drying. In this case, however, the product is not sintered (Ebnesajjad, 2000). A schematic diagram summarizing P T F E paste extrusion is shown i n F i g . 1.4 for a wire coating process (DuPont , 1994). E a c h step o f the paste extrusion process is critical i n determining the final product properties. However , various operating variables, such as lubricant content i n the paste, preforming pressure and duration, and the various die design parameters, affect the process and product i n ways that have yet to be investigated. A l though the process o f paste extrusion has been commercial ized, the fundamental mechanisms underlying the process are not fully understood. Since its inception, litde work has been done o n the subject. A detailed study o f the experimental and theoretical aspects o f P T F E paste extrusion is, therefore, necessary and w i l l be the main focus o f the present work. P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 1 - F U N D A M E N T A L S 9 PTFE Fine Powder Resin A: Removing drum from cold storage T Blending C: Preforming AW D: Paying off E: Extruding conductor J J J F: Vaporizing J: Taking up wire H: Cooling I: Spark test Fig. 1.4 Schematic diagram of P T F E paste extrusion for a wire coating process (DuPont, 1997). P A S T E E X T R U S I O N O F P O L V l ' E T R / X F L U O R O E ' n - I Y L E N E F I N E P O W D E R RESINS C H A P T E R 2: P A S T E E X T R U S I O N : G E N E R A L R E V I E W 2.1 Introduction Paste extrusion has been widely used as a fabrication technique for many useful objects, f rom everyday products such as toothpaste, penci l leads, cosmetic pencils, animal foodstuffs, and food flavorings, to less c o m m o n products such as ceramic components, catalyst supports, bricks, seal tapes, P T F E wires and cables (Benbow and Bridgwater, 1993; Ebnesajjad, 2000). Surprisingly, the subject o f paste extrusion does not seem to have been investigated systematically before. It has been presented previously either i n entirely empirical terms or as an extension o f mol ten polymer flow and extrusion. Furthermore, as far as P T F E paste extrusion is concerned, previous work is particularly scanty. In this chapter, some o f the previous studies o n paste extrusion, particularly those related to P T F E paste extrusion, are reviewed. 2.2 Experimental Aspects of Paste Extrusion Paste is essentially a mixture o f l iquid and solid particles. D u r i n g paste processing, the l iquid i n the paste matrix serves primarily as a lubricant. In P T F E paste extrusion, the l iquid has the additional function o f protecting the solid resin particles f rom being mechanically damaged pr ior to extrusion (as discussed i n Chapter 1, P T F E fine powder resins are highly susceptible to mechanical damage above the transition temperature o f 19°C) (Mazur, 1995; Ebnesajjad, 2000). B y fill ing the spaces between the particles wi th an incompressible l iquid rather than air, the paste w i l l become resistant to compressive loading, without increasing the interparticle contact area. Th is also reduces the adhesive forces between particles, particularly i n the case o f P T F E paste, since the interfacial tension for polymer-on-lubricant is much less than for polymer-on-air. This allows particles to rearrange more easily i n response to mechanical forces, without being deformed (Mazur, 1995). Consequently, the particles w i l l remain isotropic i n nature after compression, a condi t ion that is not attainable wi th a dry powder (Mazur, 1995). 10 C H A P T E R 2 - P A S T E E X T R U S I O N : G E N E R A L R E V I E W 11 T h e lubricant concentration i n a paste mixture generally varies. B e n b o w and Bridgwater (1993) have reported typical lubricant concentrations o f 35 v o l . % to 50 v o l . % , depending o n the physical properties o f the solid component , such as the particle shape and size distribution. I f the solid component o f the paste is considered as being made up o f close-packed spherical particles o f uniform size, the volume o f the v o i d spaces i n the paste matrix can then be calculated to be approximately 47 % o f the total paste volume. F o r a typical P T F E paste wi th a lubricant density o f approximately 0.75 g / c m 3 , this is equivalent to about 23 wt.%. However , P T F E paste extrusion is typically performed wi th a lower lubricant concentration, ranging between 16 wt .% to 25 wt .% (Ebnesajjad, 2000). Possibly, this has to do wi th the P T F E resins being deformable, wh ich results i n the effective v o i d volume i n the paste matrix being m u c h smaller than calculated after compression. Perhaps a more accurate representation o f the solid - lubricant arrangement i n the P T F E paste matrix is the adsorption o f the lubricant onto the surface o f the resin particles to form a thin layer o f lubricating f i lm, rather than as a v o i d filler. The amount and properties o f the lubricant critically affect the extrusion process. F o r example, based o n industrial experience, M a z u r (1995) has reported that increasing the amount o f the lubricant above some critical value results i n an extrudate that is too soft to retain its shape. O n the other hand, i f an inadequate amount o f lubricant is used, the extrudates tend to be rough and irregular. T h e viscosity o f the lubricant has also been found to have an effect o n the quality o f the paste. T h e use o f a more viscous l iquid as a lubricant has been found to result i n a less uniform mixture (Benbow and Bridgwater, 1993). K i m et al. (1997) studied the fabrication o f Y b a 2 C u 3 0 7 . 6 - A g composite super-conducting wires by a paste extrusion technique and found that, i f the viscosity o f the paste is too low, the paste may not extrude into a continuous body, or that microcracks could be developed during the drying process after extrusion. A n excellent general review on the subject o f paste extrusion is contained i n the book co-authored by B e n b o w and Bridgwater (1993). T h e authors have discussed various aspects o f paste extrusion, including the overall logic behind the design procedure for paste fabrication, extrudate surface defects and swell, lubricant migration, and the methods o f data analysis for paste extrusion. The authors also discuss the various types o f extruders that can P A S T E E X T R U S I O N O F P O L Y l ' E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 2 - P A S T E E X T R U S I O N : G E N E R A L R E V I E W 12 be used for the extrusion process, namely the rotating extruder, the ram extruder, and the screw extruder, along wi th their practical implications. In addition, comparisons are made between paste extrusion and other techniques related to powder processing. However , most o f these discussions are val id only for the extrusion o f r igid pastes, i.e. where the solid particles do not deform or change morphological ly during the extrusion process, such as alumina-based pastes. I n P T F E paste extrusion, however, where the resin particles undergo a morphological change during processing, it is not possible to use a screw extruder as a processing device, as w i l l be discussed later i n Chapter 5. Nevertheless, several conclusions from the work are wor th reviewing, since they are generally val id for most paste systems, including P T F E pastes. F o r example, B e n b o w and Bridgwater (1993) have found that the transient extrusion pressure profile f rom a ram extruder typically rises steeply at the beginning o f the extrusion process. After reaching a max imum, the pressure usually decreases gradually to a m i n i m u m until a zone o f dried and highly compacted paste (static zone) is formed around the die entrance. Further extrusion i n the presence o f such a static zone was found to cause the pressure to rise again. T h e authors point out that, at the point when the extrusion pressure begins to initially rise, the piston is i n contact wi th the paste i n the barrel for the first time. T h e authors also suggest that the m i n i m u m pressure level during the extrusion process be considered as the steady-state pressure level for analysis purposes. They claim that this is the point when the friction o n the barrel wal l is at its m i n i m u m . A similar description o f the pressure curve has also been reported by Malamataris and Rees (1993). F r o m the experimental results obtained in this work, it has been found that the above description o f the pressure response curve is applicable to P T F E paste systems as wel l , although the initial pressure rise appears more like a peak i n the case o f P T F E paste extrusion. A l s o , dur ing the steady-state extrusion o f P T F E paste, the pressure response appears to be more level than that described by B e n b o w and Bridgwater (1993), al lowing it to be more easily identified for the purpose o f experimental analysis. Further discussion o f this general aspect o f P T F E paste extrusion is given i n Chapter 5. B e n b o w and Bridgwater (1993) have also suggested a more convenient experimental extrusion procedure, i n wh ich a single batch o f paste is extruded at different speeds (during a P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 2 - P A S T E E X T R U S I O N : G E N E R A L R E V I E W 13 single experimental run) to obtain the steady-state pressure level corresponding to each o f those speeds. T h e authors, however, cautioned that this should only be carried out when there is sufficient confidence that a static zone has not formed near the die entrance. This is difficult i n the case o f P T F E paste extrusion, as w i l l be discussed i n Chapter 5. In paste extrusion, lubricant migration is an important phenomenon. It is caused by the movement o f the mobile l iquid phase i n the paste matrix that eventually results i n its non-uni form distribution i n the mixture (Mazur, 1995). Th is effect is enhanced wi th time, especially i n the presence o f high extrusion pressure. Sol id particle size distribution, shape and orientation during flow have been thought to influence the extent o f lubricant migration during paste extrusion (Benbow and Bridgwater, 1993), although no k n o w n quantitative analysis has been performed previously. It is an indication that excessive lubricant migration has occurred during extrusion when droplets o f l iquid accumulate at the die exit, and when the pressure at a lower extrusion rate is found to be higher than that at a higher rate (Benbow and Bridgwater, 1993). T h e later and progressive rise o f the extrusion pressure wi th time, as discussed above wi th reference to the pressure response f rom the ram extruder, also indicates that lubricant is being filtered out o f the bulk paste. Several authors have suggested using finer powder, or blending fine and coarse powders i n the paste mixture to reduce the extent o f lubricant migration during extrusion (Benbow and Bridgwater, 1993). It should be noted, however, that particle size also affects the porosity o f the extrudates. A method has been described by B e n b o w et al. (1987), wh ich enables the predict ion o f extrudate pore structure from particle size distribution. Lubricant migration also occurs during preforming. In studying the leftover ceramic paste after extrusion through a die having a sudden contraction, Burbidge et al. (1995) found that a conically-shaped stagnant zone was consistently formed around the die entry. F o r pastes that had been preformed pr ior to the extrusion, the size o f this artifact was larger i n size. This implies that preforming promotes lubricant migration, especially when too high a preforming pressure is used. However , inadequate pressure wi l l yield a weak product, due to insufficient removal o f entrapped air i n the paste matrix (Mazur, 1995; Ebnesajjad, 2000). A detailed study on the effects o f preforming o n paste quality is discussed i n Chapter 6 for several P T F E paste systems. P A S T E E X T R U S I O N O F POLYIK1TIAFLUOROETI-IYLENE F I N E P O W D E R RESINS C H A P T E R 2 - P A S T E E X T R U S I O N : G E N E R A L R E V I E W 14 Anothe r c o m m o n phenomenon i n paste extrusion is extrudate surface fracture. A l t h o u g h surface defects i n polymer melt extrudates have been the subject o f a large number o f studies (see, for example, Tordel la , 1956; Ramamurthy, 1986; K a l i k a and D e n n , 1987; Hatzikir iakos and Dealy, 1991), very little w o r k has been done o n the origin and elimination o f surface defects o f extruded pastes. Furthermore, there is no basis to presume similar mechanisms for fracture i n melt and paste extrusions. A general introduct ion o n surface defects i n paste extrusion has been provided by B e n b o w and Bridgwater (1993). T h e authors reported that the type and severity o f surface defects are very m u c h dependent o n paste formulation, die design and operating conditions. T o reduce the severity o f extrudate surface defects, the authors suggested the use o f extrusion dies wi th longer length to diameter ratios and smaller entrance angles. Increasing the lubricant concentration i n the paste mixture and reducing the extrusion rate were also found to improve extrudate appearance. A more comprehensive study o n the subject has been conducted by D o m a n t i (1998) and reported i n the author's doctoral dissertation. Numerous other studies have been carried out o n the extrusion o f various paste systems (see, for example, Harr i son et al, 1986; P o p o v i c h et al, 1997; Burbidge et al, 1998; W i l d m a n et al, 1999). However , work o n P T F E paste extrusion is particularly l imited. In a recent reference source for fluoropolymers, Ebnesajjad (2000) has summarized various experimental aspects o f P T F E paste extrusion. The author has discussed the general properties o f P T F E , including P T F E fine powder resin and its processing behavior. Appropr ia te selection criteria for fine powder resins and lubricants for specific applications have also been suggested, along wi th their appropriate processing conditions. T h e process o f P T F E paste extrusion has been described i n detail for various applications, such as wire coating and the fabrication o f tapes and tubes. T h e general practical implications o f many processing variables, such as die design, lubricant concentration and resin properties, have also been assessed qualitatively. M o s t importantly, the author has provided a discussion o n fibrillation as a characteristic defining phenomenon i n P T F E paste extrusion. A detailed discussion o n fibrillation can also be found i n a w o r k by M a z u r (1995). T h e first published report on the creation o f fibrils during P T F E paste extrusion is probably by Lewis and Winchester (1953). The authors noticed that, during the extrusion process, the morphology o f the P T F E resin changed from spherically shaped individual P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 2 - P A S T E E X T R U S I O N : G E N E R A L R E V I E W 15 particles to an extrudate consisting o f particles that are connected by fibrils o f submicrometer size. However , the authors provided no further detail o n the subject, particularly as far as the mechanism o f fibril creation, and the role o f operating conditions i n affecting the fibril and hence, extrudate quality are concerned. In the same work, the authors discussed the qualitative effects o f extrusion pressure, temperature and die design o n the extrudate flow rate using a constant pressure ram extruder. The authors also investigated the flow pattern o f paste at the die entry region, by performing visualization experiments, wh ich are repeated i n this w o r k (see Chapter 4) for the purpose o f further analysis. The i r results were later confirmed by Snelling and L o n t z (1960) and summarized by M a z u r (1995). In other earlier work, L o n t z and Happold t (1952) conducted a general study o n the dispersion properties o f P T F E powder. Th is was then fol lowed by a study o n the extrusion properties o f lubricated P T F E resin, i n wh ich the method o f preparing lubricated composit ions for a wire coating process was investigated (Lontz et al., 1952). It can be said that, although numerous studies o n paste extrusion have been carried out, the process o f P T F E paste extrusion, i n particular, has yet to be completely investigated. Furthermore, most o f the available published works o n P T F E paste extrusion are based o n industrial experiences that tend to be empirical i n nature. The present work is aimed at a systematic study o f the process o f P T F E paste extrusion. 2.3 Mathematical Modeling of Paste Flow The behavior o f any material under flow can be generally described as viscous, elastic, plastic, or any combinat ion thereof. T h e subject o f flow rheology has been discussed i n great detail elsewhere, and it wou ld be inappropriate to repeat all o f that information here. Mathematically, numerous constitutive models have been developed for flows o f viscoelastic materials, such as polymer melts (see, for example, Larson , 1998), and for flows involv ing solids under plastic deformation, where they are considered as ideal plastic materials (see, for example, H o f f m a n and Sachs, 1953), or as elastic-plastic materials that exhibit strain hardening (see, for example, Davis and D o k o s , 1994), as i n the case o f metal forming or wire drawing. However , little work has been devoted to the f low model ing o f paste as an elasto-viscoplastic material (this wi l l be discussed i n Chapter 8). E v e n wi th the available P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T I - I Y L E N E F I N E P O W D E R RESINS C H A P T E R 2 - P A S T E E X T R U S I O N : G E N E R A L R E V I E W 16 mathematical models, significant modifications are often necessary to improve their predictions o f the experimental data. B e n b o w and Bridgwater (1993) have proposed empirical constitutive relationships to describe the general behavior o f paste flow. These were incorporated into their derivation o f an analytical equation that predicts the effects o f die design and extrusion conditions o n the steady-state extrusion pressure. Depend ing o n various simplifications, the resulting equation may have four or six parameters that need to be determined experimentally. The authors have provided experimental verification o f their mode l predictions, using various paste systems. The fits are generally acceptable, and the solution can be generated without m u c h computational effort (see also B e n b o w et al, 1987; B e n b o w et al, 1989). However , the equation fails to predict the effect o f die entrance angle o n the steady-state extrusion pressure, as wi l l be elaborated i n Chapter 8. It has also been found that, due to its empirical nature, modifications having a theoretical basis are difficult to incorporate into the final equation. Dora iswamy et al. (1991) developed a non-linear rheological mode l for concentrated pastes, wh ich takes into account the elastic, viscous and yielding behavior o f such materials. A key feature o f the mode l is the incorporat ion o f a recoverable strain term. T h e mode l consists o f parameters wh ich can be determined solely from dynamic data. The authors claimed this as being advantageous, since these data are easily accessible by performing experiments using a parallel-plate rheometer. The authors have also shown that a correlation exists between steady shear viscosity and complex dynamic viscosity for materials that exhibit yield stress, similar to the C o x - M e r z rule for polymer melts. B y defining a new term, called the "effective shear rate", the authors were able to show the effectiveness o f the steady shear viscosity predictions from the more accessible dynamic data. However , i n the present work, experiments using a parallel-plate rheometer wi th P T F E pastes were found to be difficult and non-reproducible. It was also found that the results obtained were not useful, due to the significant slip that occurs between the paste and the rheometer plates. In another work, Louge (1996) has proposed a two-phase model for paste flow, taking into consideration phenomena such as l iquid migration, particle segregation along the walls, shear dilatancy and induced localization. The model also includes a Darcy-l ike interaction P A S T E E X T R U S I O N O F P O L Y T E T O A F L U O R O E T I T Y L E N E F I N E P O W D E R RESINS C H A P T E R 2 - P A S T E E X T R U S I O N : G E N E R A L R E V I E W 17 force and a non-linear viscous dissipation term. The author has shown that the model reduces algebraically to a p rob lem invo lv ing only one velocity field, one pressure field and the particle volume fraction. Solut ion to the model , however, requires significant computational effort. Furthermore, the mode l was derived based o n a r igid paste system, and hence, is not applicable to P T F E pastes. A simple and experimentally more consistent model to describe the flow o f P T F E pastes has been proposed by Snelling and L o n t z (1960). T h e authors consider the max imum shear stress experienced by the paste during extrusion as being made up o f two components: one proport ional to the m a x i m u m shear rate (strain hardening component) and one proport ional to the strain rate (viscous component). Th is is essentially a modif ied K e l v i n constitutive relation (Hoffman and Sachs, 1953). B y performing visualization experiments, the authors found that the flow o f P T F E paste i n the die conical zone can be described reasonably wel l by considering the "radial f low" hypothesis. Th is hypothesis assumes that all paste particles at the same radial distance from the die apex move towards the die apex at the same velocity (see Chapter 8). Based o n this hypothesis, an expression for the pressure drop was derived, which was i n good agreement with the experimental data. However , the equation does not take into account the frictional force on the die wal l , wh i ch becomes more important when a tapered die having a small entrance angle is used. A l s o , the analysis provided by the authors does not account for the pressure drop across the capillary length o f the die that follows the entrance region. In this work, the equation proposed by the authors is rederived, i n order to include the effects that had been previously neglected. There are several other previous studies that have attempted to mathematically describe the flow o f pastes i n general, such as that by H o r r o b i n and Nedderman (1998), who used the slip line analysis to predict the upper and lower boundary to the entrance pressure drop associated wi th the flow o f a perfectly elastic paste (see also H o r r o b i n , 1999). Other works include those by K u d o (1960), Kobayashi and T h o m s e n (1965), J i r i et al. (1989) and A d a m s et al. (1997). However , these solutions are generally only partially val id for P T F E paste. The fact that P T F E fine powder resin does not qualify as a rigid, or perfectly elastic material, and that it is easily fibrillated, requires P T F E paste to be categorized into a unique material class o f its o w n and analyzed differendy. P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS CHAPTER 3: SCOPE OF WORK 3.1 Introduction A l t h o u g h P T F E paste extrusion is o f a great commercial interest, little is k n o w n o f its mechanisms. This limits the ability to predict the effects o f various operating variables o n the outcome o f the extrusion process. Consequently, process opt imizat ion is often made through experimental trials, instead o f the more efficient model ing approach. Previous work o n general paste extrusion o n ceramics, alumina-based pastes, foodstuffs, etc. provided little understanding o f the P T F E paste extrusion process. T h e properties o f P T F E resins have made their rheology unique and P T F E paste extrusion fundamentally different f rom other more typical paste extrusion processes. It is tempting to describe the rheology o f P T F E pastes during extrusion as similar to that o f polymer melts. However , this is certainly not valid, as w i l l be clear f rom this work. Mathematically, previous work o n the development o f constitutive equations to mode l the f low o f polymer melts is also o f only l imited relevance to the P T F E paste system. This is because both solid and l iquid phases are present i n P T F E pastes, wi th the solid particles being fibrillated during extrusion. Thus , rheological consideration has to be extended to other systems as wel l , encompassing highly filled suspensions and even mol ten metals, rather than l imit ing it to polymer melt systems. 3.2 Thesis Objectives This research project is fundamental i n nature. F r o m this work , it is hoped that a clearer insight into the nature o f P T F E paste flow, its relation to the rheology and processing o f P T F E pastes, and the properties o f P T F E paste extrudates w i l l be gained. T h e objectives o f this work can be summarized as follows: 1. T o study the various experimental aspects o f P T F E paste extrusion, such as preforming, paste preparation, and the extrusion process itself, i n order to identify important 18 C H A F F E R 3 - S C O P E O F W O R K 19 operating variables i n P T F E paste extrusion and possibly, provide recommendations to improve the current commercia l procedure. 2. T o study the flow mechanism and morphological development o f P T F E paste during extrusion by means o f scanning electron microscopy ( S E M ) , differential scanning calorimetry (DSC) , and visualization experiments. 3 . T o understand the mechanism o f fibrillation i n P T F E paste extrusion and its effects o n extrudate properties. 4. T o determine the effects o f die design, resin molecular structure, and processing conditions o n the rheology o f P T F E pastes and the properties o f P T F E paste extrudates. 5. T o develop a mathematical mode l that is capable o f describing the rheological behavior o f P T F E pastes during extrusion. 3.3 Thesis Organization T h e first chapter o f this thesis provides fundamental information o n P T F E as wel l as its chemistry, commercial product ion and applications. The various fabrication techniques for P T F E resins are also reviewed i n Chapter 1, wi th particular emphasis on the paste extrusion process. Fo l lowing this, important literature o n paste extrusion, particularly those studies that pertain to the experimental and theoretical aspects o f P T F E paste extrusion, is reviewed i n Chapter 2. The detailed objectives o f this work are described i n this chapter. Th is is followed, i n Chapter 4, by descriptions o f the experimental apparatus and procedures used i n the present study. In Chapter 5, the general characteristics o f P T F E paste extrusion are discussed, wi th reference to the general shape o f the transient extrusion pressure profiles and the extrusion procedure. Exper imental findings on the preforming behavior o f P T F E pastes are described i n Chapter 6. Chapter 7 focuses on the flow mechanism o f P T F E pastes during extrusion and the theory o f fibrillation, as deduced from experiments invo lv ing S E M , D S C and flow visualization. The rheology o f P T F E pastes is discussed i n Chapter 8. In the same chapter, a one-dimensional mathematical model to describe the flow o f P T F E pastes is proposed and derived. The mode l predictions are illustrated i n separate subsections, along wi th their experimental verification. In Chapter 9, the effects o f various operating variables P A S T E E X T R U S I O N O F P O L Y l K f R A F L U O R O E T I I Y L E N E F I N E P O W D E R RESINS C H A P T E R 3 - S C O P E O F W O R K 20 o n the properties o f P T F E paste extrudates are discussed, wi th reference to the quality and quantity o f the fibrils formed during extrusion. Final ly, the thesis is concluded i n Chapter 10, with a general summary o f the experimental findings, the knowledge contributions resulting from this work , and some recommendations for future work. F o r ease o f reference, various symbols used throughout this thesis are listed i n a separate nomenclature section fol lowing Chapter 10. P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS CHAPTER 4: EXPERIMENTAL WORK 4.1 Introduction This chapter describes the experimental equipment and procedures used to study the paste extrusion o f P T F E fine powder resins i n this work. A capillary rheometer, an excellent mimic o f the industrial ram extruder, is an indispensable tool i n a study such as this. Detai led descriptions o f the capillary rheometers used i n the present work are given i n this chapter, along wi th the various terms associated wi th the extrusion die design. Other equipment and procedures used to determine the properties o f the extrudates are also described. This chapter also discusses the method used i n the visualization experiment. Since paste preparation is a crucial step i n P T F E paste extrusion, it is dealt wi th separately. Finally, one section o f the chapter is devoted to the descriptions o f the physical and molecular properties o f the fine powder resins, and o f the properties o f the lubricants used i n this work. 4.2 Experimental Equipment 4.2.1 Preforming and Extrusion Equipment Preforming was normally done before extrusion using a capillary rheometer wi th a blank die i n place, as wi l l be discussed later. However , for the purpose o f investigating the preforming behavior o f the different P T F E pastes, a separate unit was built. T h e preforming unit consisted o f an a luminum pipe with polished interior surface (Average Roughness, Ra, ~ 0.6 L l m ) and a steel plug, resembling a blank die. T h e unit was assembled under the capillary rheometer load frame, i n place o f the rheometer barrel, as shown i n F i g 4.1. Since the inside diameter o f the preforming unit (see F ig . 4.1) was larger than the rheometer pis ton diameter, the piston head needed to be replaced wi th a custom-made unit (it was required that the preforming unit diameter be sufficiently large to facilitate later analysis). T h e simple and light design o f the preforming unit, without the bulkiness and the weight associated wi th the rheometer barrel, allowed the removal o f the fragile preform for analysis without much effort. Th is reduced the chances o f breaking/cracking the preform during handling. A t the same time, by incorporating it into the capillary rheometer load frame, it was possible to take 21 C H A P T E R 4 - E X P E R I M E N T A L W O R K 22 advantage of the rheometer load cell and motor drive, and to monitor and adjust the preforming pressure precisely, through the available computer control board. Piston Head Sintered PTFE Preforming Unit (Aluminum Pipe) Capillary Rheometer Test Frame Steel Plug Fig. 4.1 Schematic diagram of the preforming unit used in this work. It is essentially the test frame of an Instron capillary rheometer. Capillary rheometers were used for the extrusion experiments, with the set-up shown in Fig. 4.2. The unit is comprised of a barrel, equipped with heating bands and temperature controllers, a piston, and a load frame, complete with a load cell and motor drive, as discussed above with reference to Fig. 4.1. The piston moves at a constant speed, as specified through a computer control board. A data acquisition board allows the experimental results, vi% extrusion force versus distance, to be recorded automatically and stored in a computer. During data analysis, the extrusion pressure was obtained by dividing the force required to drive the piston by the cross-sectional area of the barrel. The extrusion die was assembled into the lower end of the barrel as shown in Fig. 4.2. Stainless steel dies of various designs, each having an average surface roughness of approximately 0.02 \im, were used. For the tapered die shown in the figure, the design variables of interest were the die entrance angle (a), the die reduction (contraction) ratio (R), and the die aspect ( L / D J ratio. The die reduction ratio was defined as the ratio of the cross-P A S T E E X T R U S I O N O F POLYTETRAFLUOROE1TTYLENE F I N E P O W D E R RESINS C H A P T E R 4 - E X P E R I M E N T A L W O R K 23 sectional areas before and after the contraction zone (i.e. R = (DjDf). F o r the purpose o f the visualization experiments, special dies were designed that al lowed them to be disassembled into two halves, as shown i n F i g . 4.3. Electrical Heaters Die entrance (contraction) zone Die capillary zone-red Capillary Die Fig. 4.2 Schematic diagram of Instron capillary rheometer. The unit was assembled under the test frame shown in Fig. 4.1 for paste extrusion. A blank die was used in place of the capillary die during preforming. Fig. 4.3 (a) Die assembly for the 25.4 mm diameter capillary rheometer, (b) visualization die design. P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T I T Y L E N E F I N E P O W D E R RESINS C H A F F E R 4 - E X P E R I M E N T A L W O R K 24 Experiments were carried out at the rheology laboratory o f the Universi ty o f Bri t ish C o l u m b i a (Rheolab), using an Instron capillary rheometer, equipped wi th two interchangeable barrels having diameters o f 9.525 m m and 25.4 m m . T h e unit allowed a m a x i m u m piston speed o f 4.23 m m / s and a max imum force o f 22.3 k N . Dies o f various designs were available, including a blank die wi th OC = 90° and Da = 0. T h e bigger diameter barrel (see F ig . 4.3) was generally used for the visualization experiments. Some experiments were also performed at D u P o n t Fluoroproducts , Wi lming ton , D E , using a similar unit that allowed a max imum piston speed o f 8.3 m m / s and a m a x i m u m force o f 245 k N , wi th a barrel diameter o f 21.8 m m . 4.2.2 Other Apparatus Other major experimental equipment used i n this work included an Instron universal testing machine to determine the mechanical (tensile) properties o f the paste extrudates, differential scanning calorimeter (DSC) to determine the thermal properties o f the fine powder resins before and after extrusion, and a micro-Raman spectrometer to determine the degree o f orientation o f the fibrils created during the extrusion process. T h e use o f Raman spectroscopy to determine the molecular orientation i n drawn polymer samples has been reported previously in the literature (Voyiatzis et al., 1996; Andr ikopou los et al, 1998; De lmede et al, 1999). It involves the projections o f excited polarized Raman spectra onto a drawn polymer sample wi th the polarization geometry parallel and perpendicular to the drawing axis o f the sample. T h e intensity ratios o f certain Raman scattering bands between the two polarization geometries w i l l then indicate the degree o f molecular orientation i n the sample (a ratio o f unity indicates isotropicity and hence, no preferred molecular orientation). F o l l o w i n g this procedure, Raman spectroscopy was used i n this work i n an attempt to quantify the degree o f fibril orientation i n the paste extrudates. T h e backscattering geometry used for the Raman spectroscopy is depicted i n F ig . 4.4. T h e set-up o f the Raman spectrometer was as follows. T h e Raman spectra were excited with the linearly polarized 514.5 n m line o f an air-cooled A r + laser (Spectra-Physics 163-A42). A narrow-bandpass interference filter was used for the elimination o f the laser plasma lines. The excitation beam was directed to the sample compartment through a properly modulated metallurgical microscope (Olympus B H S M - B H 2 ) . T h e microscope was P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 4 - E X P E R I M E N T A L W O R K 25 used for the delivery of the excitation laser beam onto the sample and for the collection of the backscattered light through a beamsplitter and the objective lens adapted to the aperture of the microscope. The focusing objective was a long working distance (8 mm) 50x/0.55 Olympus lens. The spectra were obtained using a ~1 mW laser beam focused on the specimen for a total integration time of 30 s. The Raman scattered radiation was focused on the slit of a single monochromator after being passed through a notch holographic filter (HFN-514-1.0, Kaiser Optical Systems) for removal of elastic Rayleigh scattering rejection. The Raman system employed was the T-64000 model of Jobin Yvon (ISA — Horiba group). Dispersion and detection of the Raman photons were done by a 600-grooves/mm grating and a 2D CCD detector (operating at 140K), respectively. The spectral slit width was approximately 10 cm"1. The incident beam polarization was selected by an optical rotator (90°). A dichroic sheet polariser analyzed the scattered radiation, and a half-wave plate was used after the polariser whenever needed in order to ensure the same maximum polarization response from the grating. Al l spectra were corrected to take into account the beam splitter's response in the polarization of the incident and scattered radiation. The total response of the system was checked using CC14 as a reference (Voyiatzis, 2000). A A X, (to Monochromator) ' > Analyzer Beam Splitter Polarization Geometry Perpendicular to Extrusion Direction ^ Polarization Geometry Parallel to Extrusion Direction Incident Laser Beam X (Extrusion Direction) <—Paste Extrudate Fig. 4.4 Coord ina te axes for the Raman backscattering configurat ion. P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 4 - E X P E R I M E N T A L W O R K 26 4.3 Materials 4.3.1 PTFE Fine Powder Resins Five grades o f P T F E fine powder resins were available for testing i n this work. T h e resins were supplied by D u P o n t Fluoroproducts , and have the properties shown i n Table 4.1. The resin particles are generally spherical i n shape, wi th a uni form size distribution, as a result o f the polymerizat ion technique employed to produce these resins ( K i m et al, 1999a; K i m et al, 1999b; Ebnesajjad, 2001). The average particle diameter reported i n Table 4.1 are size based, and were obtained by performing laser light scattering measurements. Three o f the resins had a homopolymer structure wi th different molecular weights, and two had a slight degree o f branching due to the incorporat ion o f another perfluorinated monomer . T h e S S G data o f the resins are also included i n Table 4.1 to al low a quantitative comparison o f the resin molecular weight for resins i n the same molecular structure group. Generally, however, the molecular weight o f a P T F E resin can be inferred direcdy from its melt creep viscosity, regardless o f the molecular structure (see Chapter 1). The S S G and melt creep viscosity o f the resins were determined according to A S T M D4895-98 and U S Patent 3,819,594, respectively. The resins were kept refrigerated throughout the course o f this work to ensure a temperature lower than the P T F E transition temperature o f 19°C. Table 4.1 Properties of P T F E fine powder resins studied in this work. Relative magnitude of the resin molecular weight can be inferred proportionally from the melt creep viscosity data. Resin Relative M„ Avg. Dia. (pm) SSG Melt Creep Viscosity (Pa.s) Copolymer series: 1 L o w 0.209 2.157 1.6 x l O 9 2 High 0.204 2.153 2.1 x 10 9 Homopolymer series: 3 L o w 0.177 2.220 1.2 x 10 1 0 4 Medium 0.216 2.185 2.8 x 10 1 0 5 High 0.263 2.154 3.2 x 10 1 0 4.3.2 Lubricants A n isoparaffinic compound under the trade name o f I S O P A R G® was used as the lubricant (supplied by E x x o n M o b i l Chemicals). Its properties, as provided by the P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E l l - T Y L E N E F I N E P O W D E R RESINS C H A P T E R 4 - E X P E R I M E N T A L W O R K 2 7 manufacturer, are summarized in Table 4.2. For the visualization experiments, black dispersed pigment for PTFE extrusion supplied by Colorant-Chromatics was used in a mixture with the ISOPAR G® solvent as the lubricant. Table 4.2 Properties of ISOPAR G® solvent (Exxon Corp., 1994). ISOPAR G® Test Method Solvency Kauri-Butanol V alue Aniline Point, °C Solubility Parameter 28 83 7.3 ASTMD1133 A S T M D611 Calculated Volatility Flash Point, °C Distillation, °C Initial Boiling Pt 50% Dry Point Vapor Pressure, kPa, 38°C 41 160 166 174 1.87 A S T M D56 A S T M D86 A S T M D2879 General Specific Gravity, 15°C/T5°C Density, g/cm 3 Color, Saybolt Viscosity, Pa.s, 25°C Auto-Ignition Temp., °C Bromine Index 0.747 0.744 +30 1.00 293 <10 A S T M D1250 Calculated ASTMD156 A S T M D445 A S T M D2155 A S T M D2710 Composition, mass% Saturates Aromatic s 100 0.01 Mass Spectrometer U V Absorbance Purity, ppm Acids Chlorides Nitrogen Peroxides Sulfur None <1 <1 Trace 1 Exxon Method Exxon Method Exxon Method Exxon Method Exxon Method Surface Properties Surface Tension, mN/m, 25°C Interfacial Tension with water, mN /m, 25°C Demulsibility 23.5 51.6 Excellent duNuoy A S T M D971 Exxon Method 4.4 Experimental Procedure 4.4.1 Paste Preparation Paste was prepared by mixing PTFE fine powder resin with the lubricant in a desired mass proportion at a temperature lower than 19°C. Mixing was carried out below the PTFE transition temperature to ensure that the resin was not damaged prior to extrusion. The P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T I - I Y L E N E F I N E P O W D E R RESINS C H A P T E R 4 - E X P E R I M E N T A L W O R K 28 transfer o f the resin i n and out o f the containers was facilitated by pour ing the resin, instead o f scooping it, to further ensure the physical integrity o f the resin. T h e mix ing container was then placed i n a horizontal ro l l mixer that rotated at 15 r p m for approximately twenty minutes. The resulting mixture (paste) was aged at r o o m temperature for 24 hours before the start o f an extrusion experiment to al low more uni form wetting o f the resin particles by the lubricant. Unused paste was discarded after 10 days and replaced by a new mixture, since the lubricant concentration could have decreased wi th time. 4.4.2 Preforming In order to study the preforming behavior o f P T F E pastes, preforming was carried out using the unit described above wi th reference to F ig . 4.1. T h e paste inside the preforming unit was compressed wi th a pis ton driven by the Instron capillary rheometer. Since the computer controller o n the rheometer allowed the load cell to move only at a constant speed, the preforming procedure required the setting o f the pis ton speed at an initial value o f 0.85 m m / s and manually decreasing it as the desired pressure approached. T o maintain a constant pressure for a per iod o f time, the speed was adjusted accordingly, sometimes to as l o w as 0.0021 m m / s . The speed was never set to zero since the paste w o u l d relax and pressure decrease immediately upon the stopping o f the piston. T o eliminate paste leakage due to back pressure, a sintered P T F E rod was placed i n the preforming unit between the paste and the piston. T o determine the variation i n density and lubricant concentration along the preform, sections o f approximately 2.54 c m i n length were sliced and immediately weighed. These were dried i n an oven at approximately 90°C until the weights were constant wi th in ± 0.5 mg. F r o m the difference between the initial (wet) and final (dry) weights o f the various slices, lubricant concentration along the preform was determined. T h e density o f each slice was determined from its weight after drying and its corresponding volume. In the radial direction, the variation o f density and lubricant concentration is not expected to have a major practical impact, since i n the process o f paste extrusion, the preform is extruded through a converging die, where its diameter is appreciably reduced, making any effect due to radial variation negligible. P A S T E E X T R U S I O N O F P O I A f l T i T R A F L U O R O E T I T V L E N E F I N E P O W D E R RESINS C H A P T E R 4 - E X P E R I M E N T A L W O R K 29 In the rheological study, preforming was done using the Instron capillary rheometer wi th the barrel and a blank die i n place (see F ig . 4.2). The procedure was similar to that discussed above, wi th the preforming pressure set to 2 M P a for approximately 30 s. U p o n the complet ion o f preforming, the blank die was replaced by a tapered die and paste extrusion then proceeded. 4.4.3 Paste Extrusion and Visualisation Experiments Paste extrusion was carried out using an Instron capillary rheometer at a constant pis ton speed. T h e temperature o f the barrel was set as desired through the temperature controllers. A sintered P T F E rod was placed between the paste and the piston to eliminate paste leakage due to backpressure. T h e extrudates that were produced were collected for later analysis. Visual izat ion experiments were performed to provide insights about the flow pattern o f the paste i n the die conical zone. Experiments were carried out using the 25.4 m m diameter barrel, for ease o f analysis, along wi th its o w n set o f dies (see F ig . 4.3). The experimental procedure was similar to that discussed above. In this case, however, a second (colored) paste was prepared wi th the lubricant being a 1:1 mixture o f I S O P A R G® and the black dispersed pigment. The rheometer barrel was filled alternately wi th the colored and uncolored pastes, followed by preforming and extrusion. After the extrusion had reached steady state, the piston was stopped, and the die was removed and disassembled to recover the paste wi th in the contraction area. Subsequently, the paste was cut i n hal f axially i n order to observe the deformation pattern developed therein. 4.4.4 Extrudate Analysis Pr ior to analysis, extrudates were dried i n a vacuum oven at approximately 60°C until their weights remained approximately unchanged wi th further heating. T h e dried extrudates were tested for their tensile strength according to A S T M D l 7 1 0 - 9 6 ( P T F E rods). The thermal properties o f the extrudates were determined using a D S C according to A S T M D3418-82. Raman spectroscopy was also performed o n the extrudates at two polarization geometries (parallel and perpendicular to the extrusion direction) i n an attempt to describe quantitatively the degree o f fibril orientation i n the extrudates. P A S T E E X T R U S I O N O F P O L Y T E l * R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS CHAPTER 5: GENERAL CHARACTERISTICS OF PTFE PASTE EXTRUSION 5.1 Introduction A l t h o u g h a ram extruder can be used to study the rheological behavior o f both polymer melts and P T F E pastes, the pressure responses obtained from the extruder are characteristically different for the two types o f materials. N o t surprisingly, the physical significance o f each response is also different. Th is is largely due to the fact that both solid and l iquid phases are present i n the paste matrix, while only a l iquid phase is present i n the polymer melt system. T h e general industrial procedure for the extrusion process itself is different for the two types o f materials. F o r example, while it is possible to conduct a continuous extrusion o f polymer melts, P T F E paste extrusion has to be carried out i n a batch fashion, for reasons that w i l l be discussed below. In this chapter, the general characteristics o f P T F E paste extrusion are discussed wi th respect to the extrusion pressure response obtained from the ram extruder (capillary rheometer) and the general industrial procedure for the process. Th is w i l l provide an insight into the macroscopic flow behavior o f P T F E pastes dur ing extrusion. 5.2 Characteristics of Extrusion Pressure Curves A typical transient pressure response for P T F E paste extrusion is shown i n F ig . 5.1. Generally, the pressure curve can be classified into three distinct regions, w h i c h are identified as regions I, II, and III i n the figure. Similar pressure curves are usually obtained for other paste extrusion processes (Benbow and Bridgwater, 1993). T h e pressure peak i n region I has been commonly attributed to the initial wetting o f the extrusion die or the initial filling o f the die conical zone by the paste (McAdams , 2000). In region II , the paste flow is essentially at steady state. I n region III, the gradual rise i n pressure has been thought to be a manifestation o f non-uniform lubricant concentration, due to the l iquid phase migration i n the paste matrix (Benbow and Bridgwater, 1993; Burbidge et al., 1995). It is noted that these variations 30 C H A P T E R 5 - G E N E R A L C H A R A C T E R I S T I C S O F P T F E P A S T E E X T R U S I O N 3 1 i n the extrusion pressure may result i n the non-uniformity o f extrudate properties, as w i l l be discussed i n subsequent chapters. CD 3 i X UJ Distance in Barrel Fig. 5.1 Typical transient extrusion pressure response from a ram extruder during P T F E paste extrusion. Whi le it is clear that region II is the steady-state region o f the extrusion profile, it is also reasonable to consider region III as the extrusion zone o f a drier paste that results i n the gradual increase i n the extrusion pressure. However , preliminary experiments have indicated that region I is not entirely due to the wetting o f the extrusion die by the paste. Th is is apparent from F ig . 5.2, w h i c h shows the transient experimental results o f two extrusion runs using resin 2. In case (i), preforming was done using a blank die having a flat surface (CC = 90°) at a preforming pressure o f 2 M P a . In case (ii), preforming was done at the same pressure, but wi th a tapered extrusion die i n place, to ensure that the die conical zone is filled and hence, wetted by the paste prior to extrusion. It can be seen from F ig . 5.2, however, that the latter procedure does not eliminate the pressure peak. This raises doubt about whether the mechanism suggested for the pressure peak is indeed valid. P A S T E E X T R U S I O N O F P O L Y T E ' 1 ' R A F L U O R O E I T I Y L E N E F I N E P O W D E R RESINS C H A P T E R 5 - G E N E R A L C H A R A C T E R I S T I C S O F P T F E P A S T E E X T R U S I O N 32 25 20 h Resin 2 + 18 wt.% ISOPAR G, 35°C R = 56:1, a = 45°, UDa = 20, Q = 30.2 mm3/s Preforming with blank die (a = 90°) Preforming with extrusion die (a = 45°) _ L 4 6 8 10 Distance in Barrel (cm) 12 14 Fig. 5.2 Extrusion pressure responses with preforms obtained using different procedures. It can be seen that there is little effect on the magnitude of the pressure peak. F ig . 5.3 shows the results obtained for an experiment i n wh ich extrusion was stopped twice after reaching steady state. D u r i n g the first stop, the extrusion die was removed and reassembled immediately, before al lowing the extrusion to continue until the second steady state was reached. D u r i n g the second stop, the extrusion die was removed and the conical part o f the partially extruded preform, wh ich was molded to the shape o f the die, was removed to obtain a flat surface. The extrusion was then allowed to proceed until complet ion. It can be seen that, after the first and second stops, the peaks reappeared with a similar magnitude (peaks 2 and 3). B o t h were considerably lower than that o f the initial one (peak 1). I f the pressure peak is due to the initial wetting o f the die by the paste, then such a peak should not reappear after the first stop. However , it should reappear after the second stop, wi th a magnitude similar to that o f peak 1. A s can be seen from Fig . 5.3, this is clearly not the case. In a similar but separate experiment, cutting the conical part o f the partially extruded preform followed by wip ing the die surface free o f lubricant, d id not cause the reappearing pressure peak to have the same magnitude as that o f the initial one. P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 5 - G E N E R A L C H A R A C T E R I S T I C S O F P T F E P A S T E E X T R U S I O N 33 F i g . 5.3 Extrusion pressure responses obtained by stopping the extrusion process twice. During the first stop, the die was removed from its assembly and immediately reattached. During the second stop, the conical part of the partially extruded preform that conformed to the shape of the die was cut to result in a flat surface. T o determine i f the pressure peak is due to the elastic nature o f the paste, experiments were conducted in wh ich the extrusion process was stopped after reaching steady state. T h e n the paste was allowed to relax without disassembling the die, before proceeding further. The results are shown i n F ig . 5.4. It can be seen that al lowing the paste to relax for a short time (less than 10 minutes) caused the peak to reappear i n subsequent extrusions, although with m u c h smaller magnitudes. However , when the paste was allowed to relax for a sufficient length o f time (45 minutes), the peak d id not reappear. This implies that forming the paste into the conical shape o f the die for a sufficiendy long time causes the paste to lose its elastic memory (fading memory effect), and the pressure peak to be eliminated in subsequent extrusions. Th is explains the high pressure peak normally observed i n region I and the fact that its magnitude is irreproducible i n subsequent extrusions. That is, a virgin paste, exhibit ing full elastic memory, w i l l extrude wi th a high pressure peak at the beginning o f the process. However , due to the fading memory effect, it is not possible to reproduce the magnitude o f the first peak i n subsequent extrusions o f the same paste, without disassembling the die. It is worthwhile not ing that the steady-state extrusion pressure P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T K I Y L E N E F I N E P O W D E R RESINS C H A P T E R 5 - G E N E R A L CHARACTERISTICS O F P T F E PASTE E X T R U S I O N 3 4 increases wi th the waiting period between extrusions, as can be seen i n F i g . 5.4. Th is is not surprising since the paste becomes drier wi th time. 22 20 18 £ 16 ^ 14 « H O jg 1 2 CL C 10 O X LU -i—i—I-I—r-i—r-r-T-r-Resin 2 + 18 wt.% ISOPAR G, 35°C R = 56:1, a = 45°, UDa = 20, Q = 30.2 mm3/s 1min 3 min 1 0 m i n 45 mm 4 6 8 Distance in Barrel (cm) 10 12 Fig. 5.4 Extrusion pressure responses obtained by stopping the extrusion process a number of times, for various waiting periods (see figure). The results show the fading memory effect of P T F E paste. Further experiments were conducted i n wh ich the waiting periods between subsequent extrusions were longer. F o r these experiments, the 25.4 m m diameter barrel was used in order to facilitate the removal o f the partially extruded paste from the rheometer. The partially extruded paste was kept i n the refrigerator for approximately 48 hours before further extrusion. T h e experimental results are shown i n F ig . 5.5 (since this set o f experiments was conducted using the bigger diameter barrel, response (ii) is included for comparison purposes). It can be seen that the elastic memory o f the paste appears to be reversible. A s discussed previously, the pressure peak did not reappear i n a subsequent extrusion after a waiting period o f 30 minutes (response (ii)). However , after removing the partially extruded paste from the die and extrusion assembly for 48 hours, the pressure peak reappeared, wi th a magnitude similar to that o f the original peak (response (iii)). F r o m these experiments, it can be concluded that the pressure peak i n region I o f the extrusion pressure PASTE E X T R U S I O N O F P O L Y T E T R A F L U O R O I i F H Y L E N E F INE P O W D E R RESINS C H A P T E R 5 - G E N E R A L C H A R A C T E R I S T I C S O F P T F E P A S T E E X T R U S I O N 35 profile is more likely to be due to the elastic nature o f the paste, as opposed to the initial wetting or fill ing o f the die conical zone by the paste. 0 5 10 15 Distance in Barrel (cm) F i g . 5.5 Extrusion pressure responses obtained with longer waiting periods between re-extrusions. Experiments were conducted using the 25.4 mm diameter barrel. The results indicate that the fading memory effect is in fact reversible. 5.3 General Procedural Characteristics That both l iquid and solid phases are present i n the paste matrix, and that the solid resin particles are highly susceptible to mechanical damage, affect the general procedure wi th wh ich P T F E paste extrusion can be carried out. Variations i n the general extrusion procedure are usually very l imited, i f possible at all. F o r example, F ig . 5.6 shows the transient pressure responses f rom two sets o f experiments performed according to different procedures. In case (i), the extrusion flow rate was initially set to 60.3 m m 3 / s (i-a). Af ter reaching a steady state, the rate was decreased to 30.2 m m 3 / s (i-b), but later, increased back to the initial rate o f 60.3 m m 3 / s (i-c). In case (ii), the cycle was reversed. The extrusion was initially carried out at 30.2 m m 3 / s (ii-a), then increased to 60.3 m m 3 / s (ii-b) and later decreased back to 30.2 m m 3 / s (ii-c). It can be seen that, i n case (i), slighdy different steady-state extrusion levels were obtained for (i-a) and (i-c), although both were performed at the P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T L I Y L E N E F I N E P O W D E R RESINS C H A P T E R 5 - G E N E R A L C H A R A C T E R I S T I C S O F P T F E P A S T E E X T R U S I O N 36 same extrusion rate. A l s o , the steady-state extrusion pressure corresponding to (i-b) was similar to that for (i-a), although (i-b) was performed at a slower rate. In case (ii), the difference i n the steady-state extrusion pressures between (ii-a) and (ii-c) was larger than that between (i-a) and (i-c). These results indicate that P T F E paste extrusion is an irreversible process. That is, unlike i n polymer melt processing, where intermittent change i n the rate during an extrusion run has no effect o n the steady-state pressure, the extrusion pressure o f P T F E paste is dependent o n its shear history. It can be seen from F ig . 5.6 that, where an extrusion run has exhibited a higher steady-state pressure level at some particular rate, decreasing the rate w i l l not lower pressure (comparing (i-a) to (i-b), and (ii-b) to (ii-c)), but increasing it w i l l increase the steady-state extrusion pressure to an even higher level (comparing (i-b) to (i-c) and (ii-a) to (ii-b)). Th is is simply due to the fact that the consistency o f the paste changes wi th time and the application o f pressure. Longer times and higher extrusion pressures provide the necessary dr iving force for causing the lubricant to migrate out o f the paste matrix, leaving the paste remaining i n the extrusion barrel drier. Clearly, this effect is irreversible and w i l l have a significant influence o n the product properties, as wi l l be discussed i n later chapters. Resin 2 + 18 wt.% ISOPAR G, 35°C R = 156:1, a = 45°, L/D = 20 to CL 3 CO CO cu CL c o CO LU 35 30 25 20 15 _ I I i i i 1 i i i i 1 i i i | r i i , | i '- I \ (ii-b) "' " i : (ii-a) Q = 30.2mm 3/s - . I. . . i i i i i i Q = 60.3 mrrvVs i . Q = 30.2 mm3/s 8 10 12 Distance in Barrel (cm) F i g . 5.6 Effects of changing extrusion rate during a single extrusion run. P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 5 - G E N E R A L C H A R A C T E R I S T I C S O F P T F E P A S T E E X T R U S I O N 37 In order to improve the efficiency o f the P T F E paste extrusion process, modifications should be made to the current industrial procedure that involves a batch-wise processing technique, to al low the process to be carried out i n a more continuous fashion. However , preliminary experiments have proven this to be difficult. Firstly, since P T F E fine powder resins are highly susceptible to mechanical damage, it is not possible for P T F E paste extrusion to be carried out using a screw extruder, wi th wh ich a fully continuous process can easily be attained. T h e shear created by the rotating action o f the screw w i l l damage the resins by prematurely creating fibrils and at the same time, breaking them, causing the extrusion pressure to be excessively high. The damaged paste w i l l eventually c log the barrel, making it impossible to obtain an extrudate. Therefore, it is only possible to perform P T F E paste extrusion using a ram extruder. In order to make the process more continuous, the procedure w i l l then have to be modif ied to al low continuous feeding o f preforms into the extrusion barrel o f the ram extruder. Preliminary experiments were conducted to investigate the practical feasibility o f this procedure. T h e results are shown i n F ig . 5.7. 50 40 CO Q. a> 30 3 ° 20 in 5 >9 10 1 1 1 1 i 1 1 ' 1 i 1 1 1 1 i 1 1 1 1 i 1 1 1 1 Resin 2 + 18 wt.% ISOPAR G, 35°C R = 156:1, a = 45°, L/D = 20, Q = 30.2 mm3/s Extrusion stopped and 1st new preform inserted 1 st new preform reaches the die 2nd new preform reaches the die Extrusion stopped and t | 2nd n^ w preform ipserted | f 10 15 20 Distance in Barrel (cm) 25 30 F i g . 5.7 Effects of continuous extrusion using individually prepared preforms. It can be seen that when the beginning o f each new preform was being extruded, the extrusion pressure peaked. This indicates that different preforms do not continuously weld P A S T E E X T R U S I O N O F P O L V F E ' I ' R A F L U O R O E T I T Y L E N E F I N E P O W D E R RESINS C H A P T E R 5 - G E N E R A ] . C H A R A C T E R I S T I C S O F P T F E P A S T E E X T R U S I O N 38 to each other. Since the variation i n pressure w i l l consequently affect the extrudate properties, the proposed procedure essentially provides no improvement over the current commercial process, and therefore, has to be considered as practically unattractive. 5.4 Conclusions D u e to the presence o f both solid and l iquid phases i n the P T F E paste matrix, the extrusion pressure responses o f P T F E pastes are characteristically peculiar. T h e transient pressure profile for P T F E paste extrusions can be distinctly classified into three regions. T h e first region includes a pressure peak, wh ich was found to be due to the elastic memory o f the paste, rather than due to the initial wetting or filling o f the die conical zone by the paste. T h e second and third regions correspond to steady-state paste f low and the extrusion o f a drier paste, respectively. Variat ions i n the procedure for P T F E paste extrusion were found to be very l imited, i f possible at all. Th is is largely due to the nature o f the paste mixture, and the fact that P T F E fine powder resins are highly sensitive to mechanical damage. In an attempt to improve the efficiency o f the process by making it more continuous, it was found that different preforms do not continuously weld to each other during extrusion. Ex t rus ion o f continuously fed preforms into the extrusion barrel has resulted i n a pressure response that periodically peaked when the beginning o f each new preform was being extruded. Th i s makes the proposed modif icat ion to the procedure commercially impractical. P A S T E E X T R U S I O N O F P O L Y T E T R / \ F L U O R O E T H Y L E N E F I N E P O W D E R RESINS CHAPTER 6: PREFORMING BEHAVIOR OF PTFE PASTES 6.1 Introduction A s discussed i n Chapter 5, the extrusion pressure typically rises steeply at the beginning o f a paste extrusion process, before reaching a peak (region I) and settling to a lower steady-state level (region II). N e a r the end o f the extrusion, the pressure tends to rise slowly but continuously (region III). These variations i n extrusion pressure wi l l correspondingly affect the quality o f the extrudate. In the wire coating process, for example, it has been found that large variation i n the extrusion pressure can result i n the periodic failure o f the dielectrical spark test (McAdams , 2000). It was found that a variation i n the extrusion pressure could also occur i n region II, due to the uneven densification o f paste during preforming. This necessitates a study o n the preforming behavior o f P T F E pastes. T h e similar subject o f soil consolidation has been o f m u c h interest to soil engineers and, accordingly, much work has been done i n this area (see, for example, Craig, 1997). T h e subject o f l iquid migration through the soil has also been o f theoretical interest i n soil mechanics. In this study, experiments were conducted to investigate the preforming behavior o f P T F E pastes by determining the effects o f pressure, resin molecular weight and preforming duration, on the preform density and lubricant distribution i n the preform, using the homopolymer resins listed i n Table 4.1. The experimental findings o f this investigation are summarized and discussed i n this chapter: firstly, i n relation to the density variation i n the preform and secondly, to the extent o f lubricant migration during preforming. A n empirical relationship that relates resin S S G and preforming pressure to the effective length o f the preform (that is the length at which density begins deviating) is proposed. In addition, the effect o f applying pressure o n both ends o f the m o l d during preforming is also discussed, before the chapter is concluded i n Section 6.7. 6.2 Effect of Preforming on Extrusion Pressure Figure 6.1 plots the extrusion pressure profiles obtained using the capillary rheometer under the same extrusion conditions, wi th the preforms produced at the same pressure but 39 C H A P T E R 6 - P R E F O R M I N G B E H A V I O R O F P T F E P A S T E S 40 handled differendy. In case (i), the top o f the preform was extruded first, that is, the preform was turned upside d o w n i n the barrel. In case (ii), the top o f the preform was extruded last. In case (iii), preforming was performed o n both ends o f the paste. It can be seen that the paste near the end at wh ich pressure was applied during preforming, i.e., the top por t ion o f the preform, extrudes at a lower pressure than that o f the other end. This is shown as an increase i n pressure during the course o f extrusion for case (i) and a decrease i n pressure for case (ii). W h e n preforming is performed o n both ends, the extrusion pressure i n region II appears to be more uniform. 120 100 80 CO 0. 2 60 CO CD 40 20 i i i n ' T r i " i " i ' l ' l „ 11/ K Case (i) f [Top in Preform = Bottom in Barrel| Case (iii) | Preformed from Both Sides| Case(ii) |Top in Preform = Top in Barrel| • ' ' 20 40 60 Extension (mm) 80 100 F i g . 6.1 Effect of non-uniform preform densification on the transient pressure profile during extrusion (performing pressure of 2 MPa). Visual ly , there is also an apparent difference between the top and the bo t tom portions o f the preform. This can be seen i n F ig . 6.2, where photographs o f the two ends o f a preform are shown. The compacted (top) end (Fig. 6.2 (i)) looks smoother, while the opposite end (Fig. 6.2 (ii)) looks rougher and more granular. Th is is due to radial lubricant migration and densification that are more pronounced i n the top part o f the preform. F r o m these observations, it may be concluded that the applied pressure is dissipated after a certain distance along the preform. Mac leod and Marshal l (1977) have proposed that, besides being P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 6 - P R E F O R M I N G B E H A V I O R O F P T F E PASTES 41 used to compact the paste, the applied pressure is also used to overcome frictional forces, which manifest themselves between powder particles and along the wall of the preforming unit. Hence, the paste far from the compacted surface does not experience as much pressure, unlike in a liquid system where pressure is distributed uniformly and instantly throughout. It was also found that this effect is more pronounced with paste of higher molecular weight, as will be discussed later. Fig. 6.2 Photographs showing parts of a typical PTFE preform: (i) top and (ii) bottom. The top portion is the end at which pressure is applied. 6.3 Density Studies 6.3.1 Effect of Preforming Pressure Figure 6.3 shows a typical density variation along the preform for a high and a low molecular weight resin at different levels of pressure. It can be seen that the portion of the preform near the point where pressure was applied (top) generally has a higher density. Physically, this portion also looks smoother and glossier, compared to the bottom part of the preform, which looks rougher, drier and more granular (see Fig. 6.2). This indicates a non-uniform densification of the preform. During preforming, the resin particles are highly compacted and those adjacent to the wall of the preforming unit undergo plastic deformation that results in a smooth film of deformed powder surrounding the preform. This layer keeps the rest of the resin particles P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E 1 H Y L E N E F I N E P O W D E R RESINS C H A P T E R 6 - P R E F O R M I N G B E H A V I O R O F P T F E P A S T E S 42 inside and protects them from premature mechanical damage. It has been shown previously that these particles remain spherical even after high pressure preforming (Mazur, 1995). Th is can also be seen f rom F ig . 6.4, where S E M micrographs o f the preform surface are shown at two different magnifications. The highly compressed powder forms a f i lm and appears as a dark region i n F ig . 6.4 (ii). It can be seen that the particles surrounded by this film remain spherical. It is noted that no fibrillation occurs o n the preform surface, since all particles are m o v i n g i n a p lug f low manner and no extensional or shear forces are present. 2.0 1.9 1.8 <r- 1.7 e _o D) * 1.6 in c CD <=> 1.5 1.4 1.3 1.2 0 20 40 60 80 100 120 140 Distance from Top of Preform (mm) Fig. 6.3 Density variation along preforms for the high molecular weight (open symbols) and low molecular weight (closed symbols) resins. The preforming duration was kept constant at 0.5 minute. Error bars, indicating the standard deviation of duplicate runs, are shown only for one of the plots for clarity reasons. Similar levels of reproducibility apply to the other results. T h e high pressure at the top o f the preform results i n a greater extent o f powder compact ion i n that area. This causes the displacement o f the lubricant (the mobile phase) from the spaces between particles to all directions, including the radial direction towards the wal l , causing this part o f the preform to look glossier. A s a result, the peripheral surface o f this por t ion o f the preform w i l l be more lubricated and wi l l extrude at a lower pressure, as discussed above wi th reference to F ig . 6.1. P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T F F Y L E N E F I N E P O W D E R RESINS C H A P T E R 6 - P R E F O R M I N G B E H A V I O R O F P T F E P A S T E S 43 Fig. 6.4 SEM micrographs showing a typical surface of a perform with magnifications of (i) 80x and (ii) 10,000x. In Fig. 6.3, one can see that a pressure of 0.75 MPa is insufficient to ensure relatively even densification throughout the preform. The lower part of the preform still contains significant air voids, and insufficient lubricant has been displaced radially towards the surface. This explains the higher pressure, which is generally observed with the extrusion of the lower part of the preform (see Fig. 6.1 and the text referring to Fig. 6.1). A pressure of 3 MPa seems to be the minimum pressure necessary to produce approximately 120 mm of uniformly dense preform using this particular paste (resin 5). It was found that this pressure varies depending on the molecular weight of the resin. For example, the same extent of densification can be achieved for paste of lower molecular weight (higher SSG) at a lower preforming pressure. In Fig. 6.3, the density variation along the preform made using resin 3 at 0.75 MPa is also shown for comparison. One can see that the density variation is less for resin 3 than that of resin 5 at a comparable pressure. The relative total average preform density, normalized with the corresponding SSG value to account for the different molecular weights, is plotted versus preforming pressure in Fig. 6.5. It can be seen that the total average density is generally lower for preform made of resin with lower SSG (higher molecular weight). This is especially apparent at lower preforming pressures. At higher pressures, the total average preform density becomes approximately independent of molecular weight. P A S T E E X T R U S I O N O F P O L r i ' E T R A F L U O R O E T H Y I T i N E F I N E P O W D E R RESINS C H A P T E R 6 - P R E F O R M I N G B E H A V I O R O F P T F E P A S T E S 44 CD rr 0.82 0.80 0.78 h f 0.76 CD a eg o H 0.74 h 0.72 Duration = 0.5 min. Lubricant Concentration: —o— Resin 3 Resin 4 — A - Resin 5 18wt.% J _ 0 1 2 3 4 Preforming Pressure (MPa) F i g . 6.5 Total average preform density for the three resins, normalized with respect to the individual resin SSGs, as functions of preforming pressure. These results suggest that resin particles form a highly dense and resistive layer near the top surface o f the preform that dissipates most o f the extrusion pressure. F o r paste o f higher molecular weight (lower S S G ) , possibly due to the harder (see discussion below) and larger resin particles, higher pressure is required to cause yielding. I f the pressure is insufficiently high, particles are not being pushed d o w n the preform, creating only a small por t ion o f dense preform near the top. Hence, pressure is not transmitted axially, and this causes a large variation o f density along the preform length. Since only a small por t ion o f the preform is densely compacted, the overall average preform density w i l l be low. A s one w o u l d expect, this becomes especially significant at lower preforming pressures. F o r paste o f lower molecular weight, the resistive layer near the top o f the preform yields at a lower pressure. Hence , at the same preforming pressure, particles are n o w more "mob i l e " along the preform. Consequently, the density variation along the preform is less, and the overall preform density is higher. A good analogy to this may be that o f pour ing particles through a conical funnel. B l o c k i n g o f the funnel opening is more likely to occur i f the particles are harder and larger, and a higher pressure force is required to remove such a blockage. Th is P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 6 - P R E F O R M I N G B E H A V I O R O F P T F E P A S T E S 45 explains the relative ease o f compact ing a paste, wh ich is composed o f a lower molecular weight (higher SSG) resin, having softer particles, as w i l l be discussed later. T o further study the bulk compression behavior o f P T F E resins, each resin o f predetermined mass was compressed i n the preforming unit without lubricant at a piston speed o f 0.21 m m / s . T h e pressure was moni tored as a function o f time. A t a particular time, the volume o f the resin i n the preforming unit and hence, the resin density, can be calculated knowing the piston speed and the total length and diameter o f the preforming unit. The results are plotted i n F ig . 6.6 as a function o f pressure. It can be seen that the bulk compressibility o f each resin is approximately the same, as indicated by the same slope at each pressure. However , there is a jump i n density at about 5.5 - 7.5 M P a , as shown i n the insert o f F ig . 6.6. The jump is more sudden i n the case o f resin 3, occurr ing at a lower pressure. In the case o f resin 5, the density increases over a larger pressure interval. This sudden jump i n density implies that the resin is yielding, where the bulk volume o f the resin suddenly gives way to the applied pressure. E cn cu Q 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 . 1 1 , 1 I I 1 I 1 1 I Pure Resin without Lubricant Resin 3 -Resin 4 -Resin 5 • • / * 1.4 •j 1.2 . i i i i | i i i i | i i i i | i i i i , j * * - j s r i . .••-~'2-'' • - Vs \ -f ) V • "l 1 1 1 1 1 1 I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1" -4 5 6 7 8 9 . i i , , , . i . . , , i , , , , i , , , , 0 20 40 60 80 100 120 140 160 Pressure (MPa) Fig. 6.6 Bulk (volume) compressibility of unlubricated resins 3, 4, and 5 at room temperature under the application of a changing pressure applied by means of a piston moving at constant speed. P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 6 - P R E F O R M I N G B E H A V I O R O F P T F E P A S T E S 46 6.3.2 Effect of Preforming Duration Preforming duration is defined as the length o f time the paste is pressed at a specified pressure. Figure 6.7 plots the density variation along the preform for the high molecular weight paste (resin 5) produced wi th a preforming pressure o f 0.75 M P a for different preforming durations. O n e can see that longer preforming duration produces less axial density variation, wh ich is as expected. T h e effect o f preforming duration o n the total average preform density is shown i n F ig . 6.8. O n e can see that, at longer times, the preform becomes denser and hence, contains less voids. Th is is interesting since it indicates that time relaxation is involved. The resistive layer o f paste near the top surface o f the preform relaxes wi th time and dissipates its normal stresses onto the adjacent layer. Th is process continues eventually distributing the applied stress to the rest o f the preform body as m u c h as the time allows. A s a result, the preform becomes denser and more uniform wi th time. 2.0 1.9 1.8 1.7 ^ 1.6 .o CD ~ 1.5 U) c Q '-I 1.3 1.2 1.1 1.0 T—I—I—I—|—I—I—I—I—|—I—I—I—I—|—1—I—I—I—f— \ •I '1 Resin 5 Lubricant Concentration = 18wt.% \ Preforming Pressure = 0.75 MPa - o — Duration = 0.5 min •o •• Duration = 5 min —A— Duration = 10 min I 20 40 60 80 100 120 Distance from Top of Preform (mm) 140 F i g . 6.7 Density variation along preforms produced at 0.75 M P a with different preforming durations for the high molecular weight resin (resin 5). T h e time relaxation behavior o f the three resins was investigated and the results are plotted i n F ig . 6.9. The pure P T F E resins were initially compressed to a pressure o f 140 M P a . T h e pressure for each resin was then monitored wi th time. It can be seen that resin 5 P A S T E E X T R U S I O N O F P O L r i T i T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 6 - P R E F O R M I N G B E H A V I O R O F P T F E P A S T E S 47 relaxes faster than resin 3. This implies that the higher molecular weight resin is harder (more elastic), behaving like a hard rubber, compared to the lower molecular weight resin that is softer (less elastic). Th is explains why it is relatively easy to preform lower molecular weight P T F E paste. £ CD Q o "aj i CL O r-1.74 1.72 1.70 1.68 1.66 1.64 1.62 1.60 1.58 1.56 i—i—I—I—I—I—i—I—l—I— I— I—I—I—I—I—I—T -Resin 5 Lubricant Concentration = 18wt.% Preforming Pressure = 0.75 MPa _ l I 1 L _1_ J I I 1_ 0 2 4 6 8 10 12 Preforming Duration at Highest Pressure (min) Fig. 6.8 Total average preform density for the high molecular weight P T F E resin as a function of preforming duration. Preforming pressure was kept constant at 0.75 MPa . 6.3.3 Effect of Lubricant Concentration Figure 6.10 depicts the variation i n preform density as a function o f initial lubricant concentration for the high molecular weight (resin 5) paste, preformed by applying a pressure o f 0.75 M P a for 0.5 minute. It can be seen that a paste having a lower lubricant concentration shows a greater density variation. Visual ly , its preform also looks drier and rougher. O n the other hand, preforms that are made up o f 20 wt .% lubricant show minimal density variation for at least 120 m m . This implies that a greater lubricant concentration allows the resin particles to slide past one another more easily, thereby making the paste more compressible. A l s o , being the continuous body i n the paste, the lubricant is able to distribute stress more uniformly. Hence , wi th more lubricant in the paste, stress is more evenly distributed, resulting i n a preform having lesser density variation. Th is can be seen P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T I - I Y L E N E F I N E P O W D E R RESINS C H A P T E R 6 - P R E F O R M I N G B E H A V I O R O F P T F E P A S T E S 48 clearly i n F ig . 6.10, wh ich also contains a plot o f the average preform density as a function o f lubricant content. 1 1 1 , 1 I . 1 i i i • • • • i 1 1 1 • Pure resin with • no lubricant - Resin 3 1 Resin 4 1 •\ Resin 5 \ 1W • _ • \ , . , . i . , , i , , . . i , , , , i . , . . 0 10 20 30 40 50 Time (min) Fig. 6.9 Relaxation of pressure with time after compression of unlubricated resins. 6.4 Lubricant Migration Studies 6.4.1 Effect of Preforming Pressure Figure 6.11 shows the lubricant concentration at various points along the preform for three pastes produced by applying different pressures for 0.5 minute. T h e initial lubricant concentration o f the pastes was 18 wt.%. It can be seen from Fig . 6.11 that, for the range o f pressures used i n this work, no significant pressure effect o n lubricant migration is observed during preforming. T h i s effect may become apparent at higher preforming pressures. However , i n view o f practical considerations, such high pressures may be o f little significance. Compar ing the profiles o f the three preforms, the lubricant concentration seems to be consistently lower i n the case o f lower molecular weight preforms. This is not surprising, considering the fact that lower molecular weight paste is easier to preform. Better densification o f preform implies that a greater amount o f lubricant wi l l be displaced towards P A S T E E X T R U S I O N O F P O E Y T E T R A F L U O R O E T I T Y L E N E F I N E P O W D E R RESINS C H A P T E R 6 - P R E F O R M I N G B E H A V I O R O F P T F E P A S T E S 49 the peripheral region and be lost onto the surface o f the preforming unit. T h e distribution o f lubricant i n the preform is also more symmetrical i n the case o f the l o w molecular weight resin (resin 3), w i th more lubricant being concentrated near the center region. Th is is consistent wi th the finding reported by Y u et al. (1999) for an alumina-starch-bentonite-water paste system. In the case o f high molecular weight P T F E (resin 5), however, lubricant concentration is much lower near the top, relative to the bot tom. This is due to better compact ion near the top region o f the preform, resulting i n greater lubricant displacement towards the surface o f the preforming unit and hence, lower lubricant concentration i n the preform body. T h e bo t tom region is not as dense and looks drier, indicating that lubricant is not displaced to the periphery but remains trapped inside the paste. Total Lubricant Content (wt.%) 14 15 16 17 18 19 20 21 1.8 1.7 1.6 | 1.5 2-c 1.4 CD Q 1.3 1.2 1.1 0 20 40 60 80 100 120 140 160 Distance from Top of Preform (mm) Fig . 6.10 Density variation along preforms (high molecular weight resin) and total average preform density as functions of initial lubricant concentration. Preforming was carried out at 0.75 MPa for 0.5 minute. This observation is intriguing. A l t h o u g h preforming pressure significantly affects the distribution o f density along the preform, it does not significantly affect the extent o f lubricant migration. A s far as lubricant migration is concerned, the effect o f molecular weight far outweighs the effect o f preforming pressure, for the range o f pressure P A S T E E X T R U S I O N O F P O L r i T i T R A F X U O R O E l T - I Y L E N E F I N E P O W D E R RESINS C H A P T E R 6 - P R E F O R M I N G B E H A V I O R O F P T F E P A S T E S 50 investigated. However , an effect o f preforming pressure o n lubricant migration should not be ruled out as it is expected to become more significant at higher pressures. 18.4 18.2 C? 18.0 o-~ 17.8 o I 17.6 c CD O o 17.4 O § 17.2 3 17.0 16.8 16.6 ~|—I—I—I—I—|—I—I—I—TT3 Initial Lubricant Concentration = 18wt.% Preforming Duration = 0.5 min. 0 Resin 3 - 0.75 MPa • Resin 3 - 1 MPa A Resin 3 - 3 MPa • Resin 4 - 0.75 MPa • Resin 4 - 1 MPa A Resin 4 - 2 MPa T Resin 4 - 3 MPa © Resin 5 - 0.75 MPa 0 Resin 5 - 1 MPa A Resin 5 - 2 MPa V Resin 5 - 3 MPa 20 40 60 80 100 Distance from Top of Preform (mm) 120 140 Fig. 6.11 Lubricant concentration along preforms produced by applying different levels of pressure for 0.5 minute. 6.4.2 Effect of Preforming Duration D u e to the presence o f the l iquid phase, the length o f time for wh ich preforming is carried out is expected to have a significant effect o n the quality o f the preform. Figure 6.12 shows the effect o f preforming duration o n lubricant distribution i n the high molecular weight preform (resin 5), produced by applying a pressure o f 0.75 M P a . O n e can see that the effect is significant. A t 0.5 minute, the lubricant concentration is relatively constant along the preform. A t longer times, a gradient o f lubricant concentration develops. D u e to the loss o f some lubricant at the wal l o f the preforming mo ld , the lubricant concentration i n the preform is generally lower than 18 wt.%. D u r i n g preforming, the pressure imposed at the top surface o f the preform provides a dr iving force, i n addition to the force o f gravity, for the lubricant to migrate downward. K e e p i n g the preforming pressure constant for a longer per iod o f time w i l l result i n a more P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS CHAPTER 6 - PREFORMING BEHAVIOR OF PTFE PASTES 51 severe lubricant migration and hence, a greater lubricant concentration gradient. Consequendy, at longer times, lubricant concentration becomes higher near the bot tom o f the preform, as shown i n F ig . 6.12. It can be seen that the bo t tom part o f the preform may have a lubricant concentration even higher than the original 18 wt .% when the preforming duration is sufficiendy long. Hence , i n such cases, the top part o f the preform w i l l extrude at a higher pressure, while the bo t tom part o f the preform may become "too wet" and extrude at a lower pressure, resulting i n an extrudate o f varying strength. (It has been previously reported that unnecessarily large lubricant concentration w i l l result i n a weak extrudate (DuPont , 1994; Ebnesajjad, 2000).) Therefore, while longer preforming duration may reduce the variation i n preform density, it significantly increases the extent o f lubricant migration. This results i n a preform wi th a gradient o f lubricant concentration, which , for practical reasons, is much less attractive. 19.5 19.0 18.5 S. c o | 18.0 c CD O 8 17-5 c CD CJ £ 17.0 16.5 16.0 0 20 40 60 80 100 120 140 Distance from Top of Preforming (mm) Fig. 6.12 Lubricant concentration along preforms produced at 0.75 M P a with different preforming durations (high molecular weight resin). 6.4.3 Effect of Eubricant Concentration The effect o f initial lubricant concentration o n lubricant migration is shown i n F ig . 6.13. It can be seen that a paste wi th greater initial lubricant concentration wi l l produce a PASTE EXTRUSION OF POLYTETRAFLUOROETFrYLENE FINE POWDER RESINS C H A P T E R 6 - P R E F O R M I N G B E H A V I O R O F P T F E P A S T E S 52 preform having a larger lubricant concentration gradient, wh ich is not surprising. Therefore, while increasing the lubricant concentration decreases the variation i n preform density, it does so by increasing the extent o f lubricant migration. It should be noted, however, that the results i n F ig . 6.13 correspond to short preforming duration o f 0.5 minute. T h e variation i n lubricant concentration along the preform is expected to be more significant when longer preform times are used. 21 20 19 c o 2 18 h c cu o c o O c CO 17 16 15 h 14 O - - O -- O - -o-Resin 5 Preforming Pressure = 0.75 MPa Preforming Duration = 0.5 min. 15wt.% Initial Concentration 18wt.% Initial Concentration 20wt.% Initial Concentration - O - O _l_ _1_ 20 40 60 80 100 120 140 Distance from Top of Preform (mm) 160 Fig. 6.13 Lubricant distribution along preforms having different initial concentrations of lubricant (high molecular weight resin). Preforming was carried out at 0.75 M P a for 0.5 minute. 6.5 Predicting the Effective Preform Length F r o m the above discussions, it is concluded that longer preforming times and higher lubricant contents increase the uniformity o f preform density. However , these conditions are practically unattractive, since they promote lubricant migration. Preforming pressure seems to be the only adjustable variable since its effect o n lubricant migration is relatively insignificant. Therefore, to minimize variation i n extrusion pressure due to the uneven densification o f the preform, it is necessary to prepare the preform by applying an adequate pressure. It was previously shown i n Section 6.2.1 that this pressure varies depending o n the molecular weight o f the resin. P A S T E E X T R U S I O N O F P O L Y F E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 6 - P R E F O R M I N G B E H A V I O R O F P T F E P A S T E S 53 Figure 6.14 shows that the length o f preform at wh ich the density starts deviating is not a function o f the total preform length. T w o experiments performed using different amounts o f paste indicated approximately the same variation o f preform density. The extent o f lubricant migration, however, is greater i n the case o f the shorter preforms, due to the fact that the same pressure is applied over a shorter distance (higher pressure gradient). 2.0 1.9 1.8 1.7 E 1-6 _o cn £ 1-5 '</) c ID 1 4 Q > A 1.3 1.2 1.1 1.0 i—i—i—i—i—r-Resin 5 Preforming Pressure = 0.75 MPa Preforming Duration = 0.5 min. -O— First Trial ~o— Second Trial with Less Material - •— Lubricant Concentration (1st Trial) - «— Lubricant Concentration (2nd Trial) _i_ 18.1 18.0 17.9 17.8 ? 1 7 7 I 17.6 £ CD i i.o o o 17.4 § 17.3 -§ 17.2 17.1 17.0 20 40 60 80 100 120 Distance from Top of Preform (mm) 140 Fig. 6.14 Density variation and lubricant distribution along preforms having different lengths. F r o m experimental results such as those plotted i n F ig . 6.3, it is possible to determine the effective lengths o f preform at wh ich the density starts deviating for the different resins at different levels o f preforming pressures. Such experimental results were analyzed and found to be a function o f resin S S G and preforming pressure, as shown i n F ig . 6.15. F o r a performing pressure between 0.75 M P a and 3 M P a , and a lubricant concentration o f 18 wt.%, the effective length o f preform (/*) can be written as a function o f S S G and preforming pressure as: / * (mm) = a.SSG.[P(MPa)]'' + c, (6.1) where I* is taken to be the length when density starts to deviate by approximately 0.05 g / c m 3 , and a, b and c are empirical constants, wh ich are found to be 857, 0.0229 and -1817, P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T I - I Y L E N E F I N E P O W D E R RESINS C H A P T E R 6 - P R E F O R M I N G B E H A V I O R O F P T F E P A S T E S 54 respectively for the homopolymer PTFE pastes studied in this work. These constants are expected to vary with the initial lubricant concentration. For the high molecular weight paste at a preforming pressure of 0.75 MPa, the effective length correlates linearly with the initial lubricant concentration in a semilog plot, as shown in Fig. 6.16. 120 i i i i i | i i i i | i i i i | i i i i | i i i i | i i i i | i i i i 0 I . . . . i i I 2.12 2.14 2.16 2.18 2.20 2.22 2.24 2.26 SSG*P 0 0 2 2 9 (MPa00229) Fig. 6.15 Length of preform of uniform density as a function of resin SSG and preforming pressure. Equation (6.1) indicates that a higher pressure is required to produce a certain length of preform of uniform density for a lower SSG or higher molecular weight resin. As discussed above, this is partially due to the hardness of the high molecular weight resin, causing it to yield at a higher pressure during compression. The larger raw particle dispersion size of the high molecular weight resin also contributes to this effect. Larger particles tend to be less mobile during compression, requiring greater pressure for densification. Although Eq. (6.1) is empirical in nature, it is useful for determining the preforming pressure that will adequately produce a certain length of uniformly dense preform. It should be noted that preform diameter may also have a significant effect on these results (Macleod P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T I - r Y L E N E F INE P O W D E R RESINS CHAPTER 6 - PREFORMING BEHAVIOR OF PTFE PASTES 55 and Marshall, 1977) but has not been considered in this discussion. This issue should be the subject of a future study. 10 2 h c o 2 '> C D Q S 10< <D o <D m C D C cu 10° L —]—l—l—l— l—|—l—l—l—I—|—l—l—l—l—|—l—l—l—l—|—l—I—l—l—["" 14 15 16 17 18 19 20 21 Initial Lubricant Concentration (wt.%) Fig. 6.16 Length of preform of uniform density as a function of initial lubricant concentration for the high molecular weight resin (resin 5). 6.6 Effect of Preforming From Both Ends Preforming by applying pressure on both ends will decrease the variation of density as shown in Fig. 6.17. However, the total variation in lubricant concentration remains approximately the same, with the maximum concentration near the middle of the preform. This finding is encouraging. It indicates that an evenly dense preform may be produced at a lower pressure by compacting the paste at both ends without sacrificing the preform quality in terms of lubricant migration. In this case, pressure was applied first on one end and subsequently on the other. Consideration of applying pressure on both ends simultaneously would certainly lower lubricant concentration gradient further. 6.7 Conclusions Preforming is a critical part of the paste extrusion process. It has been shown that preforming pressure and duration significantly affect the quality of preform. Lack of PASTE EXTRUSION OF POLYTE1TIAFLUOROETHYLENE FINE POWDER RESINS C H A P T E R 6 - P R E F O R M I N G B E H A V I O R O F P T F E P A S T E S 56 2.0 1.9 1.8 1.7 IE 1-6 o i 1-5 r> CO 8 1 4 F-1.3 1.2 F-1.1 1.0 I 1 ' 1 1 I -1—I—I—I—I—r— 0 Resin 5 Lubricant Concentration = 1 Preforming Pressure = 0.75 I Preforming Duration = 0.5 min. —O— One Sided Preforming —D— Two Sided Preforming —•— Lubricant Concentration (1 Sided) H Lubricant Concentration (2 Sided) 20 40 60 80 100 120 Distance from Top of Preform (mm) 18.3 18.2 18.1 18.0 17.9 17.8 17.7 17.6 17.5 17.4 17.3 17.2 140 Fig. 6.17 Ef fec t o f p re fo rming f rom bo th ends o n density and lubricant dis t r ibut ion (high molecular weight resin). adequate pressure w i l l result i n a preform o f non-uniform density wh ich wi l l extrude unsteadily, resulting i n an unacceptable final product. Increasing lubricant concentration and preforming time improves the uniformity o f preform density, indicating that the process o f preforming is time dependent. However , lubricant migration becomes important at longer times. This leaves preforming pressure as the only adjustable parameter. It was found that the pressure required to adequately produce a preform o f uni form density depends o n the molecular weight (standard specific gravity) o f the resin. It has been shown that lower molecular weight P T F E resin (higher standard specific gravity) has softer particles, wh ich are easier to compact. Finally, it has been shown that making a preform by compacting the paste o n both ends improves the uniformity o f preform density without sacrificing its quality i n terms o f lubricant distribution. P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R R E S I N S CHAPTER 7: MECHANISM OF PTFE PASTE FLOW 7.1 Introduction T h e flow mechanism associated wi th P T F E paste extrusion differs significantly from polymer melt flow. In paste extrusion, for example, the distribution o f the l iquid phase i n the paste matrix during high pressure extrusion is an important issue, especially when longer extrusion times are involved, wh ich may cause a non-uniformity o f the l iquid distribution (Mazur, 1995). A l s o , microscopically, solid state P T F E molecules are confined i n their crystallite and spherulite configurations, while i n polymer melt, molecules are randomly positioned, not conformed to a specific shape, and are significantly more mobile . In a number o f ways, the process o f P T F E paste extrusion is more similar to ceramic paste processing. However , since fibrillation is involved i n the mechanism o f P T F E paste flow (as w i l l be discussed later), the resulting extrudate is relatively much stronger. F o r example, i n P T F E paste extrusion, free flowing powder is processed to form a preform o f some initial yield strength and later, an extrudate o f significant strength (due to fibrillation) that is commercial ly useful, even without sintering, such as i n the case o f calendered P T F E tape (Ebnesajjad, 2000). In this chapter, the mechanism o f P T F E paste f low during extrusion is discussed. S E M , wh ich was used to "track" the morphological changes that occur during the course o f the process, provides an insight into the development o f extrudate strength from free flowing paste and hence, the formation o f fibrils. T o understand the deformation pattern involved therein, visualization experiments were performed, and the results are discussed i n a separate section. Finally, a "radial flow" hypothesis is proposed to describe the f low pattern o f the paste i n the contraction zone o f the die, where fibrillation was found to occur. 7.2 Morphological Development during PTFE Paste Extrusion T o determine the morphological changes that take place during the course o f P T F E paste extrusion, S E M was performed o n the P T F E paste before and after preforming, as wel l as after extrusion. Typica l S E M micrographs are shown i n F ig . 7.1 for resin 2. It can be 57 C H A P T E R 7 - M E C H A N I S M O F P T F E P A S T E F L O W 58 seen that the unprocessed particles o f P T F E powder are spherical i n shape and have a uniform size distribution (Fig. 7.1 (a)). After high pressure preforming, the solid particles o f the paste still retain their spherical shape (Fig. 7.1 (b)). Th is is due to the presence o f the l iquid phase that, besides acting as a lubricant, also provides a cushioning effect to the resin particles, preventing premature fibrillation from occurr ing as they slide past one another (Mazur, 1995). Af ter extrusion through a tapered die, the resin particles remain generally spherical and do not appear deformed (Fig. 7.1 (c)). However , they are n o w connected to each other by fibrils o f submicron size, wh ich account for the final strength o f the extrudate. This indicates that the creation o f fibrils occurs i n the die conical zone. Furthermore, judging from the typical thickness o f a fibril and considering the fact that the resin particles remain spherical after extrusion (indicating no permanent shape deformation), it is reasonable to exclude the possibility o f fibrils as being deformed particles. Instead, it is more likely that fibrils are collections o f molecules that have been unwound from their crystallite configurations. In F ig . 7.1 (c), one can also see a definite preferred orientation o f these fibrils i n the flow direction, wh ich may be further improved by calendering (Ebnesajjad, 2000) or using dies o f longer L/D ratio (this wi l l be discussed i n Chapter 9). 7.3 Mechanism for Fibrillation A proposed mechanism for fibrillation is shown i n F i g . 7.2. D u r i n g the extrusion process, compacted resin particles entering the die conical zone are highly compressed due to the reduction i n the flow cross-sectional area. T h e amorphous regions o f the P T F E crystallites (see F ig . 1.3) i n neighboring particles then begin to mechanically interlock, as the particles rub against one another under the application o f high pressure, while compet ing for an exit out o f the die. Th is results i n the interconnection o f adjacent particles. A t the exit o f the die, these connected particles experience an accelerated flow, during wh ich the mechanically locked crystallites are consequently unwound, creating fibrils. The particles later relax and return to their original spherical shape. It is important to note that high pressure extensional flow such as that w h i c h occurs i n the die conical zone is necessary for fibril creation. Simple shearing action o f loosely compacted paste at a relatively l o w pressure does not result i n a practically useful extent o f fibrillation. Th is can be seen i n F ig . 7.3, wh ich shows a S E M micrograph o f a P T F E paste P A S T E E X T R U S I O N O F P O L Y T E T T I A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 7 - M E C H A N I S M O F P T F E P A S T E F L O W 59 (resin 2) obtained after a simple shearing action between two circular plates, one of which rotates in a cyclic manner (parallel plate rheometer). When Fig. 7.3 and Fig. 7.1 (c) are compared, the difference in the extent of fibrillation is clear. (c) Fig. 7.1 SEM micrographs of resin 2: (a) before processing (virgin), (b) after preforming, and (c) after extrusion. Similar morphologies are observed for other pastes. To test the validity of the proposed mechanism, DSC analyses were performed on the extruded samples. Since the value of the PTFE first heat of melting, A H m l , is directly proportional to the degree of the resin crystallinity (Cavanaugh, 1999), unwinding of crystallites should result in a decrease in the A H m l value. Therefore, an extruded (i.e. fibrillated) sample is expected to have a lower A H m l value compared to that of the virgin resin. A H m l values of extrudates obtained under various extrusion conditions for different resins are compared with the corresponding A H m l values of the virgin resins in Fig. 7.4 P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 7 - M E C H A N I S M O F P T F E P A S T E F L O W 60 ( A H m ] values of virgin resins were obtained by performing differential scanning calorimetry on the corresponding dried pastes). It can be seen that fibrillation has apparendy decreased the degree of resin crystallinity, as indicated by the consistently lower A H m l values of the extruded samples. These results help validate the proposed mechanism for fibrillation discussed above. The difference in the A H m l values before and after extrusion should, therefore, provide a quantitative indication of the extent of fibrillation that has occurred during the extrusion process. This will be discussed in more detail in Chapter 9. (a) (b) (c) Fig. 7.2 Schematic diagram illustrating the proposed mechanism for fibrillation: (a) compacted resin particles enter the die conical zone, (b) resin particles are highly compressed and in contact with one another in the die conical zone, resulting in the mechanical locking of crystallites, (c) upon exiting the die, particles return to their original spherical shape, and entangled crystallites are unwound, creating fibrils that connect the particles. Fig. 7.3 SEM micrograph of resin 2 after shearing in a parallel plate rheometer. P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS CHAPTER 7 - MECHANISM OF PTFE PASTE FLOW 61 6 2 6 4 6 6 6 8 7 0 7 2 7 4 AH m 1 Before Extrusion (J/g) Fig. 7.4 Values o f the first heat of melting obtained from D S C analysis for pastes before and after extrusion under various conditions. 7.4 Pattern of Flow Deformation 7.4.1 Visualisation Experiments B y extruding alternate layers o f colored and uncolored disks o f paste as described i n Chapter 4, the pattern o f P T F E paste flow during the extrusion process can be determined. Typica l results f rom the visualization experiments are shown in F ig . 7.5. It can be seen that i n the barrel (Fig. 7.5 (a)), the paste moves i n a p lug flow manner wi th no radial velocity gradient, as indicated by the layers o f paste that remain horizontally level. I n the conical zone o f the die, radial and axial velocity variations become more apparent. It can be seen that the center particles move towards the die apex at a velocity greater than that o f the peripheral particles. Th is is more pronounced i n a die wi th a greater entrance angle. In fact, when a die wi th an abrupt contraction (Ct — 90°) is used, the paste forms a static zone at the die corner, where it becomes highly compacted and, consequendy, much drier than the flowing paste. This results i n the flow being streamlined at an effectively smaller entrance angle, imply ing PASTE EXTRUSION OF POLYTETRAFLUOROETI4YLENE FINE POWDER RESINS C H A P T E R 7 - M E C H A N I S M O F P T F E P A S T E F L O W 62 that there exists a critical entrance angle beyond which there is no further effect on the flow pattern development. (c) (d) Fig. 7.5 Velocity profiles obtained from visualization experiments: (a) in the barrel, (b) in the die conical zones with a = 30°, (c) a = 45°, and (d) a = 90°. 7.4.2 Radial Flow Hypothesis Snelling and Lontz (1960) have proposed the "Radial flow" hypothesis to describe the flow of PTFE paste in the conical zone of the die (see Chapter 2). The hypothesis assumes that paste particles at the same radial distance from the virtual apex of the die conical zone move towards the die apex at the same velocity. In other words, the hypothesis assumes the existence of a virtual spherical surface of radius r, as measured from the die apex, which spans over an angle of 2a, describing the position of the paste particles that move at the same velocity. This is illustrated in Fig. 7.6. By performing similar visualization experiments as those described above, and measuring the distances of the center and the corresponding peripheral particles from the die apex, and comparing them with the predictions obtained P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 7 - M E C H A N I S M O F P T F E P A S T E F L O W 6 3 f rom the "radial f l ow" hypothesis, Snell ing and L o n t z (1960) have shown the consistency o f this hypothesis wi th experimental observations. In support o f this claim, the positions o f the paste particles i n the die conical zone as they move towards the die exit have been tracked mathematically i n this w o r k for various (X, using the fol lowing equation: dr D — = — , (7.1) dt 27r ( l - COS a)r2 ' where Q is the volumetric f low rate o f the paste and the term 27l(1 -cosGC)/ represents the area o f the virtual spherical surface, i n accordance to the "radial f low" hypothesis. T h e results are shown i n F ig . 7.7 for a = 10°, 30° and 45°. Compar ing Figs. 7.5 (b) and 7.5 (c) wi th Figs. 7.7 (b) and 7.7 (c), respectively, one can see that there is indeed a general similarity between the experimental and predicted patterns. In F ig . 7.7 (a), one can also see that when a is sufficiently small, the f low i n the die conical zone almost follows a p lug f low pattern. This implies that wal l slip and hence, fact ional analysis, needs to be considered when model ing the flow o f paste, especially when the die entrance angle is small. F o r a — 90°, calculation o f the velocity profile provides non-informative results, since effectively, the paste flows at a smaller angle, due to the presence o f the static zone discussed above. F r o m these calculations, it can be seen that the "radial flow" hypothesis can be validly used to describe the flow pattern o f P T F E paste i n the conical zone o f the die. Consequently, this hypothesis w i l l be used as the basis o f the paste flow mode l i n the subsequent chapter. Virtual Surface Fig. 7.6 Schematic illustration of the "Radial F low" hypothesis. The hypothesis assumes the existence of a virtual surface of radius ras measured from the die apex, on which all paste particles moving towards the apex have the same velocity. P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 7 - M E C H A N I S M O F P T F E P A S T E F L O W 64 30 Fig. 7.7 Flow patterns in the die conical zones as calculated based on "radial flow" hypothesis with Di, = 9.525 mm, Q = 75.4 mmVs, and (a) a = 10°, (b) a = 30°, and (c) a = 45°. Time increments were set arbitrarily. P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 7 - M E C H A N I S M O F P T F E P A S T E F L O W 65 7.5 Conclusions T h e mechanism o f flow involved i n P T F E paste extrusion is fundamentally different f rom polymer melt flow. It has been shown through S E M that P T F E particles are fibrillated during the paste extrusion process. Fibr i l la t ion was found to occur at the conical zone o f the die, making its design a crucial aspect o f the process. A mechanism for fibrillation has been proposed and validated using D S C . The proposed mechanism considers fibrils that connect the P T F E particles i n the extrudate as unwound crystallites, wh ich are mechanically locked to one another pr ior to exiting the die. Th rough visualization experiments, the flow patterns o f the P T F E paste i n the barrel and i n the die conical zone have also been determined. It was found that the flow i n the die conical zone closely follows a p lug flow pattern when the entrance angle is small. The flow pattern becomes more deformed as the entrance angle is increased. In addition, for a die wi th an abrupt contraction (OC — 90°), a static zone at the die corner was observed. Mathematically, it was found that the "radial flow" hypothesis is able to describe the paste flow pattern i n the die conical zone consistently. The hypothesis assumes that particles at the same radial distance from the virtual die apex move towards the die apex at the same velocity. P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS CHAPTER 8: RHEOLOGY OF PTFE PASTES 8.1 Introduction This chapter discusses the rheological behavior o f P T F E pastes during extrusion. A one-dimensional mathematical model , based o n the "radial flow" hypothesis described i n Chapter 7, is derived. The effects o f extrusion conditions, i n terms o f extrusion temperature and rate, and the lubricant concentration i n the paste, are discussed. F o l l o w i n g this, a section is devoted to discussing the effects o f die design. The model predictions are presented and discussed i n the last section. 8.2 Theoretical Considerations: Development of 1-D Mathematical Model 8.2.1 Paste Flow through a Tapered Orifice Die (L/Da — 0) Figure 8.1 shows a volume element bounded by the spherical surfaces o f radius r and (r + dr) as measured from the virtual die apex, and by four planes at the azimuthal locations o f d, 6 + dd, <p and (j) + d0. T h e "radial flow" hypothesis implies that this element w i l l flow towards the die apex, such that its bounding surfaces remain parallel to those at its previous posi t ion. Since the element does not rotate or deviate f rom its straight path, this also implies that the stresses acting o n the element are purely normal stresses. In fact, these stresses are principal stresses, wi th the radial direction and the directions normal to the four bounding planes as the principal directions. W i t h symmetry taken into account (i.e. <7U = Om - C7(j), the stress tensor can, be written as: °1 0 0 <7r 0 0 0 0 = 0 0 0 0 0 0 T h e forces acting o n the volume element are shown i n F ig . 8.2 and the force balance i n the radial direction proceeds as follows: (a) Resultant force contr ibution i n the radial direction from CT; 66 C H A P T E R 8 - R H E O L O G Y O F P T F E P A S T E S 67 -(<Jr+d(JrXr + dr)zsinO # d8+Orr2smd d<$> dd. Neglect ing differentials o f higher orders, this becomes - r 2 s i n 0 d<p dd d(7r - 2rO r s i n0 d(j) dd dr, and integrating f rom (/) = 0 to (j) = 2K and 6 = 0 to 0 — a yields - 2nr2 ( l - cosCf) dGr - 4 ^ r ( l - co sOl f t r dr. Fig. 8.1 (a) Volume element and (b) its dimensions in the conical zone of a tapered die according to "radial flow" hypothesis. (b) Resultant force contribution i n the radial direction from <7g: The forces normal to the four bounding planes produce a resultant force i n the radial direction o f magnitude (see F ig . 8.2 (b)) P A S T E E X T R U S I O N O F POLVrETRAFLUOROETHYLENE F I N E P O W D E R RESINS C H A P T E R 8 - R H E O L O G Y O F P T F E P A S T E S 68 2(o" ersin0 dr dd #) Integrating f rom (j) = 0 to 0 = 2K and 6 = 0 to 6 = a yields 4 ^ o " e r ( l - c o s a ) dr. (a r+da r)(r+dr) 2sineded(|> (b) Fig. 8.2 Force balance on volume element: (a) forces acting on volume element, (b) radial contribution from the four normal forces. (c) Fr ic t ional force contribution i n the radial direction: Cou lomb ' s law o f friction states that frictional force is proport ional to the normal force acting o n a surface with proportionality constant ,^ the coefficient o f friction. Hence, o n the die surface, the frictional force acting i n the radial direction is (see F ig . 8.3) f<Jer sin a d(j) dr. Integrating from <p = 0 to (j) = 2K yields P A S T E E X T R U S I O N O F P O L Y F E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 8 - R H E O L O G Y O F P T F E P A S T E S 69 2KfOer sin a dr. rsinad<t>dr cosadr (*) (b) Fig. 8.3 Normal and frictional forces acting on surface element (a) top view, (b) side view. Equilibrium condition requires that the sum of these forces vanishes, i.e. 27Tr 2(l-cosa) dar - 47Cr(\ - cosa)cTr dr + 4/rcT er(l-cosa) dr + 27tpersma dr = 0. f sina (8.2) By letting B- — -and N, - <7e- 0„ and rearranging, the following equation is 2(1 - cos a) obtained dr r (8.3) The term N, is the familiar first normal stress difference in polymer melt analysis (Dealy and Wissbrun, 1990). In order to solve the above differential equation, a relationship describing the first normal stress difference for the solid-liquid (paste) system in question is, therefore, required. P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E ' l H Y I . E N E F I N E P O W D E R RESINS C H A P T E R 8 - R H E O L O G Y O F P T F E P A S T E S 7 0 F o r an ideally plastic material, Saint-Venant's theory o f plastic flow states that, at the incipience o f yielding, AT, = O~0, where <70 is the initial yield stress o f the material (Saint-Venant , 1870). However , for a completely plastic flow to occur wi th in an elasto-viscoplastic material, Nf has to sufficiently exceed CT so as to overcome the initial yield stress, the elastic flow zone and any viscous resistance that may develop during the flow (Hoffman and Sachs, 1953; Chakrabarty, 1998). T h e generalized Newton ' s law for viscous flow states that a = T]'s and Hooke ' s law o f elasticity establishes the relationship O = Ee, where T] and E are the viscosity coefficient and Young ' s modulus, respectively, and s a n d eare the logarithmic strain and strain rate tensors, respectively. C o m b i n i n g the two laws, K e l v i n proposes the fol lowing stress-strain relationship for a visco-elastic material (Hoffman and Sachs, 1953): a = Ee+rie . (8.4) U s i n g the above relation, the term OQ - OR adopts the form o f cr e - t r r = E(ee -er)+r](ee -er), (8.5) where £ e — £ r = £ „ —£, and £ e — £ r = £ ] I — £ 7 are the max imum strain, 7 m a x , and the max imum strain rate, / m a x , respectively. T h e term m a x i m u m strain was first introduced by L u d w i k (1909), w h o realized that N, should be a unique function o f / m a x . L u d w i k was also credited for the modif ied Hooke ' s law expression that takes the final form o f a power law equation CT = C e " , (8.6) where C i s Young ' s modulus when n - 1. D u e to the presence o f both the l iquid and solid phases i n the P T F E paste system, it is necessary to consider P T F E paste as an elasto-viscoplastic material. T o mode l its flow, the expression suggested by Snelling and L o n t z (1960) has been adopted. It is essentially Ke lv in ' s stress-strain relation (Eq . (8.4)), wi th modifications that are similarly employed i n P A S T E E X T R U S I O N O F P O L Y I T i T R A F L U O R O E l T T Y L E N E F I N E P O W D E R RESINS C H A P T E R 8 - R H E O L O G Y O F P T F E P A S T E S 71 the L u d w i k power law model (Eq . (8.6)) for the elastic (strain hardening) term, and the generalized power law model for the viscous resistance term. The resulting expression for the first normal stress difference, Nt, is, then, as follows: °e-°r=C7mJ' +mmJ- (8-7) A similar stress-strain relation has also been employed by Dav i s and D o k o s (1944) i n their analysis o f metal wire drawing, although without the inclusion o f the viscous resistance term. T o account for the initial yield stress, an additional term may be included o n the right hand side o f E q . (8.7). However , this term is expected to be negligible compared to the other terms, as indicated by the initial strength o f the preform, wh ich is much less than that o f the extrudate. N o w , a volume element at a distance rb f rom the virtual die apex experiences a total linear strain o f rdd-r,dd r e e = -£— = 1 (8-8) rbdd rb i n the ^-direction as it travels to a new posi t ion r away f rom the die apex (see F ig . 8.1). In terms o f logarithmic strain, this can be written as ed = l n ( l + e e) = l n — (8.9) r,, or dr d£e = — . (8.10) Provid ing the volume o f the element remains constant, it can be proven that d£r+d£e+d£,=0 (8.11) P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS CHAPTER 8 - RHEOLOGY OF PTFE PASTES 72 and since dee = de^, then dr d£r=-2d£e=-2— (8.12) r and hence, •dr = 3 ^ r, = -3 In-*-. (8.13) r T h e max imum strain rate can then be expressed as max max ^ = ^ = 3 — . (8.14) dt rdt Consider ing the "radial flow" hypothesis, it is noted that the local velocity o f a particle o n the spherical surface o f radius r f rom the die apex is given by E q . (7.1) dr Volumetric Flowrate Q dt Area of Surface 27r(l — cos Of)r^ (7.1) Hence , the max imum strain rate is 3jg 7 m a x = - ft X 3 - ( 8 - 1 5 ) 2 ^ ( 1 - cos a)r Substituting into E q . (8.3), the force balance becomes dr where do, CT 2N, ( l + B) r -2B^ = ^ - , (8.3) PASTE EXTRUSION OF POLYTETRAFLUOROETHYLENE FINE POWDER RESINS C H A P T E R 8 - R H E O L O G Y O F P T F E PASTES 73 - N , =<x r-a 0 =C 3 In + 77 [27t(l-cosa)r3 j (8.16) Solv ing the above equation yields the following expression: - C T r = 2 ( l + r 3 > 2 B j c f (31r4/r))" -(3m + 2B) 2 B + 1 3<2 dr + 2 7 r ( l - c o s a ) f 1 A 3 m + 2 B . 2 B P + r C, (8.17) where the constant o f integration, C, is evaluated using the boundary condi t ion Or = <7r when r = r , i.e. C = - ^ - 2 ( 1 + B)jc[ (31n(r 4 /r))" *7 ( 3 » + 213) 3J2 Y V 1 A [ 2 ^ ( l - c o s a ) J 3 » + 2 B (8.18) T h e extrusion pressure can then be calculated using Eqs . (8.3) and (8.13) wi th r = rb, i.e. [ ^ 2 s i n a J ^ m ^ , r 2 B + 1 (3m + 2B) ( 1 2 j g s i n 3 q ^ ?z;(l - c o s a ) L > 3 ^ ( R B + 3 ^ - I ) (8.19) where <7ra is the stress at the die exit, R is the reduction ratio, defined as (DJDJ2, and C, T], n, m a n d / a r e material constants that have to be determined experimentally. W h e n an orifice die is used, Ora may be present at the die exit due to a pull ing force during extrudate wind-up or calendering. However , 0~ra is typically negligible and the expression for extrusion pressure can then be simplified to P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T F I Y L E N E F I N E P O W D E R RESINS CHAPTER 8 - RHEOLOGY OF PTFE PASTES 74 P , . =2(1 +BMC extrusion V / ( D , f B p = ^ (Mrjr))" {?>m + 2B) 2 sin Of \2Q s i n ' a ^ ( l - c o s a ) D i 3 J dr + ( R B + 3 ^ - I ) (8.20) wh ich reduces to the expression obtained by Snelling and L o n t z (1960) when B = 0 (i.e. frictionless surface). Numer ica l integration is required to solve E q . (8.20). However , for the range o f the die reduction ratio o f interest, the fo l lowing approximation can be used wi th reasonable accuracy i n E q . (8.20), allowing an analytical solution to be obtained: \n(rjr)~a(rjr)h, (8.21) where a and b are constant fitting parameters. 8.2.2 Paste Flow through a Tapered Die (L/Da * 0) T h e forces acting on a volume element i n the capillary zone are shown i n F ig . 8.4. A force balance o n the element yields / \ D2 D2 (o-z + do, ) r - 2 - - atn -f- = farMDadz 4 4 (8.22) or d<Jz _4fcxr, _4f(Nl+<7z) dz D, D. (8.23) where JV, is the previously defined first normal stress difference, wh ich is expected to be significant due to the elastic nature o f P T F E paste. A t the end o f the die conical zone (hence, at the entrance o f the die capillary zone), N, can be calculated using E q . (8.16) with r — ra. A s s u m i n g Nt to be approximately constant throughout the capillary zone o f the die, the force balance becomes PASTE EXTRUSION OF POLYTETRAFLUOROETHYLENE FINE POWDER RESINS C H A P T E R 8 - R H E O L O G Y O F P T F E P A S T E S 75 doz 4/{Nu+CJz) where dz ("\ > ( | l „ ( R ) + rj I2 J V D. (8.24) 1 2 j g s i n 3 a R 3 / 2 ^ 3 '* J n(l-cos a)Db Solving E q . (8.24) and applying the boundary condi t ion o"z = <7ZL at z — L, yields (8.25) (8.26) where OZL is the stress imposed at the exit o f the die, wh ich is typically negligible or zero, b z = L Fig. 8.4 Force balance on volume element in the die capillary zone. T h e stress present at the entrance o f the die capillary zone, OZO, is obtained from E q . .26) wi th z = 0. Since 0~Zo is essentially OM o f E q . (8.19), the final expression for the total P A S T E E X T R U S I O N O F P O L Y l ' E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 8 - R H E O L O G Y O F P T F E P A S T E S 76 extrusion pressure across a tapered die can then be obtained by substituting Ora — 0ZO into E q . (8.19), i.e. (31n(r,/r))" P . =cr. =0" RB + 2(l + BNC extrusion rb ra V / i n a ^=17^ 2si < 2B+1 *7 (3« + 2B) ^ 1 2 j 2 s i n 3 a A ; r ( l - c o s a ) D 3 ( R B + 5 ^ - I ) (8.19) where (8.27) and f3 > ( | l n ( R ) + 71 K 2 ) V 1 2 5 s i n 3 a R 3 / 2 ^ ^ • ( l - c o s a ) D / (8.28) 8.3 Effects of Extrusion Conditions The effect o f extrusion temperature o n the steady-state extrusion pressure is shown i n F ig . 8.5 for resin 2. It can be seen that paste extrusion at a higher temperature requires a greater pressure, wi th the effect being more pronounced at a higher extrusion speed. This is due to the lower viscosity o f the lubricant that causes it to be more mobile i n redistributing itself non-uniformly i n the paste matrix. This , coupled wi th the high pressure overshoot during the initial stage o f the extrusion, causes most o f the lubricant to be extruded early. T o o high a temperature also promotes l iquid evaporation. Consequently, the paste is drier and this causes the observed increase i n the steady state extrusion pressure at higher temperatures. In F ig . 8.5, one can also see that the steady-state extrusion pressure generally increases wi th an increase i n the extrusion rate (volumetric flow rate), due to the increase i n the viscous resistance o f the paste. However , at l ow extrusion rates, an initial reduction i n the extrusion pressure is observed. This can be explained by considering the residence time o f the paste i n the barrel and the die, wh ich becomes sufficiently long at l ow extrusion rates, P A S T E E X T R U S I O N O F P O L Y T E T R / \ F L U O R O E T I - I Y I . E N E F I N E P O W D E R RESINS C H A P T E R 8 - R H E O L O G Y O F P T F E P A S T E S 77 providing ample time for the lubricant to re-distribute itself non-uniformly i n the paste matrix. The consequence is an early extrusion o f most o f the lubricant, as discussed above. A s a result, the required extrusion pressure is higher than expected, wi th the effect becoming more pronounced wi th further lowering o f the extrusion rate. This explains the trend observed i n F ig . 8.5. 0 20 40 60 80 100 120 140 160 Volumetric Flow Rate (mm3/s) Fig. 8.5 The effects of temperature and extrusion speed on the steady-state pressure of paste extrusion. The effect o f lubricant concentration on extrusion pressure is depicted i n F ig . 8.6 for resin 3 (model predictions are discussed below). A s expected, increasing the lubricant concentration decreases the extrusion pressure. However , when the lubricant concentration is unnecessarily high, the extrudate becomes excessively wet and weak, sometimes not being able to hold its shape. It is the excessive local slippage between flowing particles i n the die that results i n a lesser quantity o f fibrils being formed. O n the other hand, extruding a paste wi th a lubricant concentration lower than the op t imum value w i l l result i n a higher extrusion P A S T E E X T R U S I O N O F P O L Y r E T R A F L U O R O E T I T Y L E N E F I N E P O W D E R RESINS C H A P T E R 8 - R H E O L O G Y O F P T F E P A S T E S 78 pressure that may cause fibril breakage (this w i l l be detailed i n Chapter 9). Th is implies that the extent o f fibrillation is dependent o n the lubricant concentration as well . 100 90 80 "to 70 o_ §. cu 60 i _ CO to _„ <u 50 CL S 30 20 10 0 20 30 40 50 60 70 80 Volumetric Flow Rate (mm3/s) Fig. 8.6 The effect of lubricant concentration on the steady-state extrusion pressure for resin 3. Solid lines are model predictions obtained with fitted parameters listed in Table 8.2. 8.4 Effects of Die Design T h e effects o f the die reduction ratio, the JL/D a ratio, and the die entrance angle o n the extrusion pressure are illustrated i n Figs. 8.7, 8.8, and 8.9, respectively. The model predictions are also shown i n the same figures, using the fitted values o f the parameters listed i n Table 8.1. T o obtain these parameters, it was first assumed t h a t / = 0. Preliminary fitting o f Eq. (8.19) to the experimental data indicated that this is a reasonable assumption when a is sufficiendy large (Cf > 10°). U s i n g the available data corresponding to Of > 10°, the parameters C, n, 77, and m were first determined fol lowing the method described by Snelling and L o n t z (1960). (Eq. (8.19) essentially reduces to the expression suggested by authors i f / = 0.) Af ter obtaining the values o f these parameters , /was then determined by min imiz ing the sum o f differences between the model predictions and the experimental data. L 1 1 1 1 I 1 1 1 1 I 1 1 1 1 I 1 ' 1 1 I 1 1 1 1 I 1 1 1 1 J Resin 3 + ISOPARG, 35°C R = 352:1, a = 45°, L/D. = 0 A 22 wt.% ISOPAR G j i i i I i i i i I i i i i ! i i i i I i i i i 1 i i i i P A S T E E X T R U S I O N O F P O L Y ' I ' E T R A F L U O R O E T I - r Y L E N E F I N E P O W D E R RESINS C H A P T E R 8 - R H E O L O G Y O F P T F E P A S T E S 79 100 200 300 400 500 Reduction Ratio F i g . 8.7 The effect of reduction ratio on the steady-state extrusion pressure for resins 1—4. Solid lines are model predictions using the fitted parameters listed in Table 8.1. P A S T E E X T R U S I O N O F P O L Y I K I T I A F L U O R O E T I - I Y L E N E F I N E P O W D E R RESINS C H A P T E R 8 - R H E O L O G Y O F P T F E P A S T E S 80 P A S T E E X T R U S I O N O F P O L Y T E T R / V F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 8 - R H E O L O G Y O F P T F E P A S T E S 81 Fig. 8.8 The effect of die L/Da ratio on the steady-state extrusion pressure at different reduction ratios for (a) resin 1, (b) resin 2, (c) resin 3, and (d) resin 4. Solid lines are model prediction using the fitted parameters listed in Table 8.1. P A S T E E X T R U S I O N O F P O L r f E T R A F L U O R O E T I - I Y L E N E F I N E P O W D E R RESINS C H A P T E R 8 - R H E O L O G Y O F P T F E P A S T E S 82 70 I i i i i | . i i i | i i i i | i i . i | . . i i | 10 I . . . . i . . . , , , , i , , , , i I 0 10 20 30 40 50 Entrance Angle, a (b) F i g . 8.9 The effect of die entrance angle on the steady-state extrusion pressure: (a) at different extrusion rates for resin 3, (b) at 75.4 mm 3 / s for resins 1—4. Solid lines are model predictions using the fitted parameters listed in Table 8.1. Also shown in (a) is the prediction using the Benbow-Bridgwater equation (1993). F r o m Figs. 8.7 and 8.8, respectively, it can be seen that extrusion pressure increases wi th increases i n the reduction and L/Da ratios. This is due to increased levels o f strain hardening and frictional losses, respectively. It can also be seen that extrusion pressure is significantly affected by the die reduction ratio, especially i n the case o f high molecular weight resin. Therefore, extrusion dies o f precise (low tolerance) dimensions are essential i n paste extrusion, especially when a small diameter barrel is used. In F ig . 8.9, one can see that the extrusion pressure initially decreases, and then increases, wi th increase o f the die entrance angle. T h e decrease i n the extrusion pressure at small entrance angles is similar to the trend predicted for polymer melts using the lubrication approximation (Dealy and Wissbrun, 1990). A theoretical derivation similar to that employed i n the lubrication approximation, has been used by B e n b o w and Bridgwater (1993) i n their model ing work on paste flow. W h e n the die entrance angle is sufficiently small, paste flow i n P A S T E E X T R U S I O N O F P O L Y T E T R j \ F L U O R O E T I - r Y L E N E F I N E P O W D E R RESINS C H A P T E R 8 - R H E O L O G Y O F P T F E P A S T E S 83 the die conical zone follows essentially a p lug flow pattern for the most part, as has been discussed i n the previous chapter wi th reference to F ig . 7.6 (a). T h e lubrication approximation approach is val id here. However , the lubrication approximation predicts a monotonic decrease i n the extrusion pressure wi th increasing die entrance angle, wh ich is not consistent wi th the experimental results plotted i n F ig . 8.9. Beyond a certain entrance angle, the extrusion pressure increases wi th increasing OC, as is commonly observed wi th the extrusion o f elastic solids (see, for example, H o r r o b i n and Nedderman, 1998 and the references therein). T h e failure o f the lubrication approximation is due to the assumed plug flow pattern i n the die conical zone that becomes inval id when the entrance angle is large. A s discussed before, this can be remedied by considering the "radial flow" hypothesis. Table 8.1 Values of material constants and coefficients of friction for the different pastes. Res in C ( M P a ) n T| ( M P a . s m ) m f Copolymer series: 1 0.2348 1.61 2.767 x I O 6 1.91 4.48 x 10-3 2 0.1974 1.94 4.242 x I O 5 1.75 2.03 x I O 3 Homopolymer series: 3 0.2280 1.75 1.417 x I O 2 1.04 4.02 x I O 3 4 0.1420 1.94 5.286 x I O 2 0.88 3.57 x IO-3 8.5 M o d e l P r e d i c t i o n s -F r o m Figs. 8.7, 8.8 and 8.9, one can see that the proposed one- dimensional mode l is able to predict the flow behavior o f P T F E pastes reasonably well , considering its simple form. However , it is possible to improve the mode l fit further, especially at l o w extrusion rates, by considering the dependency o f the material constants on l iquid redistribution i n the paste matrix, for wh ich additional studies are required. Compar ing the resin properties o f the homologous series i n Table 4.1 wi th the model parameters i n Table 8.1, it can be observed that the value o f n is larger for a higher molecular weight paste. This indicates that the increase i n the extrusion pressure, due to the increase i n the strain level, w i l l be relatively greater for such paste. This is also somewhat apparent from the slopes o f the curves plotted i n F ig . 8.7. Correspondingly, i n the case o f a higher molecular weight paste, the extrudate strength w i l l be more significantly affected by changes i n strain, such as changes i n the die P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T I T Y L E N E F I N E P O W D E R RESINS C H A P T E R 8 - R H E O L O G Y O F P T F E PASTES 84 reduction (contraction) ratio. A t the same strain level, however, the effect o f molecular weight o n strain hardening (and hence, o n extrudate strength) is less apparent since the strain hardening coefficient, C , decreases wi th the increase i n the resin molecular weight, therefore compensating the increase i n n. A paste o f higher molecular weight also exhibits a greater flow resistance, as indicated by the larger value o f 77. I n addition, since the value o f m is smaller, a higher molecular weight paste behaves more like an elastic-plastic material. Th is implies that its flow behavior is less dependent o n the extrusion rate. A lower molecular weight paste flow, o n the other hand, w i l l show a greater dependency o n the extrusion rate, due to the larger value o f m. This makes the assumption o f an ideal p lug flow i n the die capillary zone less appropriate for such paste, and may possibly explain the observed discrepancies between the mode l predictions and the experimental data o n the effect o f die L/Da ratio (Fig. 8.8), wh ich , as one can see, is greater i n the case o f the lower molecular weight paste. Interestingly, as shown i n Table 8.1, the coefficient o f friction, / also decreases wi th increasing molecular weight. Perhaps, this has to do more wi th the average resin particle size that affects the mechanism o f l iquid migration to the die wall , rather than the effect o f the resin molecular weight itself. Generally, however, it can be summarized that, at practical strain levels, a higher molecular weight paste extrudes at a higher extrusion pressure, as is found experimentally (see F ig . 8.7). F o r a particular paste, the relative magnitudes o f the strain hardening and viscous resistance terms depend o n the extrusion conditions and die design, wh ich is apparent from E q . (8.19). F o r example, this is shown i n F ig . 8.10 for resin 4. The total contribution o f the two terms indicates the extent o f fibrillation that has occurred during an extrusion process. A higher extrusion pressure translates into the creation o f more fibrils (this w i l l be discussed i n more detail i n Chapter 9). However , it is worthwhile not ing that too high an extrusion pressure may be detrimental to the quality o f the fibrils, causing them to break. In the more severe cases, excessive extrusion pressure may even result i n a non-continuous spurting o f periodically broken extrudates. Th i s was observed wi th the extrusion attempt with resin 5 using a die having a reduction ratio o f 352:1. A s a consequence, it was not possible to determine the mode l parameters for this resin due to the lack o f experimental data. P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T I - I Y L E N E F I N E P O W D E R RESINS C H A P T E R 8 - R H E O L O G Y O F P T F E P A S T E S 85 P A S T E E X T R U S I O N O F P O L V r E T R A F L U O R O E T F T Y L E N E F I N E P O W D E R RESINS C H A F F E R 8 - R H E O L O G Y O F P T F E P A S T E S 86 F i g . 8.10 The relative magnitudes of the strain hardening and viscous terms in E q . (8.19) as functions of (a) volumetric flow rate, (b) die entrance angle, and (c) die reduction ratio, for resin 4 with 18 wt.% I S O P A R G® at 35°C. The effect o f lubricant concentration o n the values o f the material constants has also been studied using resin 3. The results are summarized i n Table 8.2 and the model predictions are plotted i n F ig . 8.6. It can be seen that the coefficient m, wh i ch indicates the viscous dependence o f the extrusion pressure o n the extrusion rate, increases wi th increase i n lubricant concentration. This is consistent wi th the expectation o f obtaining a more viscous response wi th an increase i n the l iquid phase concentration. The resistance to flow, as indicated by the viscosity coefficient, 77, and the coefficient o f friction, / also decreases wi th increasing lubricant content, as expected. O n the other hand, a higher lubricant concentration decreases the extrusion pressure dependence o n strain, as indicated by the lower value o f n. A l t h o u g h the strain hardening coefficient, C, increases wi th increasing lubricant concentration, at a practical strain level, the overall effect is a lower extent o f strain hardening. Thus , a paste with a higher lubricant concentration wi l l extrude at a lower P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T I T Y L E N E F I N E P O W D E R RESINS C H A F F E R 8 - R H E O L O G Y O F P T F E P A S T E S 87 pressure (see F ig . 8.6), and this w i l l translate into a lower extent o f fibrillation and hence, weaker extrudates. Table 8.2 Values of material constants for resin 3 with different lubricant concentrations. Material Constant 16 wt.% Lubricant 18 wt.% Lubricant 22 wt.% Lubricant C (MPa) 3.346 x IO 4 0.2280 0.7394 n 4.97 1.75 0.53 T| (MPa.sm) 2.302 x 10 1 1.417 x IO 2 6.748 x IO-3 m 0.72 1.04 1.08 f 6.55 x 10-3 4.02 x IO-3 1.80 x IO-3 8.6 Conclusions In this chapter, the rheological behavior o f P T F E pastes is discussed. T h e effects o f extrusion conditions, i n terms o f extrusion temperature and speed, and the lubricant concentration i n the paste, on the steady-state extrusion pressure have been investigated. The effects o f die design have also been assessed quantitatively, and a simple mode l has been proposed to describe these effects. The model considers the paste as an elasto-viscoplastic material that exhibits bo th strain hardening and viscous resistance effects during flow. The model was found to be consistent wi th the experimental data and able to predict the steady-state extrusion pressure reasonably well , given the fitted parameters o f the generalized K e l v i n constitutive relation used to describe the rheological behavior o f the pastes. P A S T E E X T R U S I O N O F P O L Y T I H T I A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 9: P R O P E R T I E S O F P T F E P A S T E E X T R U D A T E S 9.1 Introduction Discussions o n fibrillation and its mechanism during P T F E paste extrusion were dealt wi th i n detail i n Chapter 7. A s may be expected, the extent o f fibrillation and the quality o f the fibrils formed during the extrusion process are significantiy affected by the resin properties, extrusion conditions and the design o f the extrusion die. These variables consequently affect the final product properties, such as the mechanical strength o f unsintered calendered tapes, the dielectric breakdown property o f wires and the stretch v o i d index o f tubes and hoses (Ebnesajjad, 2000). In this chapter, the effects o f die design, resin molecular structure and lubricant concentration on the mechanical properties o f P T F E paste extrudates, and their relations to the quantity and quality o f the fibrils formed during extrusion are discussed. 9.2 Quantitative Descriptions of Fibrillation: The Issue of Fibril Quantity and Quality In order to completely describe the state o f fibrillation i n an extrudate, it is necessary to consider both the extent o f fibrillation i n the extrudate and the quality o f the fibrils that are formed. D u r i n g paste extrusion, a significant por t ion o f the extrusion pressure is dissipated to overcome the strain hardening and viscous resistance contributions o f the paste f low (see Chapter 8). The frictional forces between individual solid particles, and between the bulk paste and the die wal l also give rise to additional pressure drop. These factors consequently translate into the creation o f fibrils, i n accordance wi th the proposed mechanism for fibrillation discussed i n Chapter 7. Therefore, the extrusion pressure can be used as a quantitative measure o f the extent o f fibrillation that occurs during extrusion. F o r example, the extrusion o f paste through a die o f larger reduction ratio requires a higher extrusion pressure and consequently, this w i l l result i n an extrudate that is more fibrillated, as w i l l be discussed later. However , it is noted that since fibrillation occurs only in the die entrance (contraction) zone, the additional pressure drop across the capillary part o f the die should not be taken into account when describing the extent o f fibrillation during extrusion. 88 C H A P T E R 9 - P R O P E R T I E S O F P T F E P A S T E E X T R U D A T E S 89 Since fibrillation effectively results i n a decrease i n resin crystallinity due to the unwinding o f crystallites, it is also theoretically possible to quantify the extent o f fibrillation i n an extrudate using D S C . In Chapter 7, it was noted that the first heat o f melt ing o f a polymer is directly proport ional to the degree o f resin crystallinity. Therefore, the difference i n the first heat o f melting o f the paste, A H m l , before and after extrusion should be indicative o f the amount o f fibrils formed during extrusion, and should, therefore, correlate wi th the extrusion pressure. In F ig . 9.1, the relative difference i n the first heat o f melt ing o f the paste before and after extrusion is plotted as a function o f the steady state extrusion pressure under various processing conditions. A m i d the scatter, it is still possible to extract a positive correlation between the extrusion pressure and the difference i n A H m l . T h e scatter can be attributed to the fact that the samples tested i n the calorimeter represent only minute parts o f the extrudates, malting the results highly localized and hence, sensitive to any local variations i n the amount o f fibrils. Nevertheless, the results do confo rm wi th the proposed mechanism for fibrillation, i n wh ich fibrils are considered as unwound crystallites, and conf i rm the significance o f extrusion pressure i n fibril formation. F r o m F ig . 9.1, one can also see that different correlations are obtained for different resins, wh ich implies that different levels o f extrusion pressure are required to fibrillate resins o f different molecular structure under identical conditions. Besides the amount o f fibrils that are present i n the extrudate, the quality o f the fibrils is an equally important issue to consider when describing the mechanical properties o f extrudates. The quality o f fibrils can be described i n terms o f their degree o f orientation and continuousness. A preliminary attempt to quantify the degree o f fibril orientation in an extrudate using Raman spectroscopy showed promising results. Figure 9.2 (a) and (b) show typical Raman spectra for unprocessed and processed powder, respectively. In the case o f unprocessed powder (Fig. 9.2 (a)), no difference i n the scattering intensity at all Raman shifts is observed between the two polarization geometry (i.e. parallel or perpendicular to the extrusion direction), indicating no preferred orientation (isotropicity). In the case o f processed powder (Fig. 9.2 (b)), however, there is a definite difference i n the scattering intensity between the two polarization geometries at major Raman shifts, particularly at 734 cm" 1 and 1383 cm ' 1 . The assignment o f Raman scattering bands for P T F E is difficult and controversial, especially those at 734 cm" 1 and 1383 cm" 1 (Bower and Maddams, 1989; P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 9 - P R O P E R T I E S O F P T F E P A S T E E X T R U D A T E S 9 0 Lehnert et al, 1997). There is general agreement that these bands are due to C - F and C - C stretching (Bower and Maddams, 1989). However , the question o f wh ich band is associated wi th wh ich stretching, is not yet resolved. Nevertheless, it was found that the Raman scattering intensity at 1383 cm" 1 is consistendy enhanced i n the polarization geometry parallel to the extrusion direction, while that at 734 cm" 1 is enhanced i n the polarization geometry perpendicular to the extrusion direction. F r o m S E M , it is evident that, for the fibrils to become oriented, the P T F E molecules have to be straightened wi th their backbones parallel to the extrusion direction (see F ig . 7.1 (c)). Therefore, it can definitely be concluded that the Raman scattering band at 1383 cm" 1 corresponds to C - C (backbone) stretching. T h e ratio o f the Raman scattering intensities at 1383 cm" 1 between the two polarization geometries (parallel to perpendicular) w i l l then provide a quantitative measure o f the preferred fibril orientation. A ratio o f unity indicates isotropicity or no preferred orientation, while a ratio greater than unity indicates a preferred orientation i n the direction parallel to the extrusion direction. x X < X < 0.10 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 -1 1 1 T - - 1 — i — I — i — i — i — V ;7o 8 / / ° / / / / y / / / • i . . . . i O Resin 2 (Copolymer) A Resin 3 (Homopolymer) • Resin 4 (Homopolymer) _J 1 I L 1 20 40 60 80 100 120 Steady State Extrusion Pressure (MPa) Fig. 9.1 Correlations between the relative differences in A H m i of pastes before and after extrusion and the steady-state extrusion pressure (under various extrusion conditions). The fact that the differences in A H m i are always positive indicates that the resin crystallinity is consistendy lower after extrusion. The lines are drawn to guide the eye. P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E U W L E N E F I N E P O W D E R RESINS C H A P T E R 9 - P R O P E R T I E S O F P T F E P A S T E E X T R U D A T E S 91 c a> t: CD CD r r -i—i—i—i—I—i—i—i—i—I—i—i—r T — I — I — I — r — I — I — r Parallel to extrusion direction Perpendicular to extrusion direction 734 cm' 1 parollal perpantHcitar 1302 crrf'l 1216 cm'' 250 500 750 1000 1250 1500 Raman Shift (cm1) (a) 1750 CD CD > CD r r -j—i—i—i—i—p Parallel to extrusion direction Perpendicular to extrusion direction 734 cm'' T pnr psndfc ular pard le I 1383 c m ' L l 250 500 750 1000 1250 1500 1750 Raman Shift (cm"') (b) Fig. 9.2 Typical Raman spectroscopy results for (a) unprocessed powder and (b) paste extrudate. P A S T E E X T R U S I O N O F P O L Y T E T O A F L U O R O E T I T Y L E N E F I N E P O W D E R RESINS C H A P T E R 9 - P R O P E R T I E S O F P T F E P A S T E E X T R U D A T E S 92 In this work, Raman spectroscopy has also been used to track the development o f fibril orientation i n the extrudate during the course o f an extrusion experiment. Th is is shown i n F ig . 9.3 for resin 4. T h e transient extrusion pressure response is also plotted i n the same figure. The degree o f fibril orientation, as shown wi th reference to the right axis, initially rises wi th the extrusion pressure. However , when the extrusion pressure is at its peak, a sudden drop i n the degree o f fibril orientation is observed. W h e n the process has reached steady state, the extrusion pressure and the degree o f fibril orientation i n the extrudate level o f f to their steady-state values. The l o w degree o f fibril orientation at the peak pressure explains the commonly observed dielectric spark test failures i n the wire coating process at approximately the same point during the extrusion (McAdams , 2000). W h e n the extrusion pressure is at its peak, the extrudate is highly accelerated out o f the extrusion die. This causes the extension and breakage o f fibrils. Consequently, the fibrils become less continuous and are more chaotic and hence, exhibit a lower degree o f preferred orientation. This results i n a non-uni form sintering and the manifestation o f voids i n the final product, causing the commonly observed spark test failures. 100 80 co CL a> 60 3 to 0) CD .2 40 U J 20 0 i | i i i i | i i i i | i i i i Resin 4 + 18wt.% ISOPAR G, 35°C a = 30°, Q = 125 mm3/s R = 336:1, L/Da = 20 i\ A I ID W W W ^ l l l l l l l | W>|^N»^>t l»*^WWII o Extrusion Pressure Raman Intensity Ratio _ L 1.75 1.70 1.65 5 CO DC >. '55 1.60 | _c c CO 1.55 £ 1.50 1.45 0 20 40 60 80 100 Distance travelled by piston in capillary rheometer (mm) Fig. 9.3 Development of fibril orientation with extrusion pressure during transient extrusion experiment. Raman intensity ratio is defined as Iparaiiei/Iperpcndicuiar at 1383 c m 1 . P A S T E E X T R U S I O N O F P O L V r E T R A F L U O R O E T I - r Y I . . E N E F I N E P O W D E R RESINS C H A P T E R 9 - P R O P E R T I E S O F P T F E P A S T E E X T R U D A T E S 93 A n industrially acceptable paste extrudate should, therefore, be sufficiently fibrillated, wi th the fibrils being mosdy unbroken (continuous) and exhibit a high degree o f orientation i n the extrusion direction. It is also important that the extent o f fibrillation and the quality o f the fibrils be uniform along the extrudate. The consequence o f these conditions is a mechanically strong extrudate. Consider ing this, the tensile strength o f an extrudate may be used as a quality indicator, since it takes into account the combined effects o f fibril quantity and quality. Furthermore, mechanical testing results are generally more reproducible wi th less variability, making extrudate tensile strength a suitable parameter for this purpose. 9 .3 Effects of Die Design 9.3.1 Effect of Die Reduction Ratio T h e effect o f die reduction ratio on the extrudate tensile strength is shown i n F i g . 9.4 for resins 2 and 4. It can be seen that increasing the die reduction ratio initially increases the tensile strength o f an extrudate. A t larger reduction ratios, the effect diminishes. In fact, the results for resin 2 indicate a small decrease i n the extrudate tensile strength at a reduction ratio o f 352:1. 7 48.6 MPa 1 Resin 2 + 18wt.% ISOPAR G, 50°C \ L/Da=20, a = 45°,Q = 30.2mm3/s -- a - Resin 4 + 18wt.% Lubricant, 35°C -UDa = 0, a = 45°, Q = 75.4 mm3/s ; 25.6 MPa —ED 62.4 MPa 0 50 100 150 200 250 300 350 400 Die Reduction Ratio Fig. 9.4 The effect of die reduction ratio on extrudate tensile strength. Also shown are the steady-state extrusion pressures corresponding to each experimental run. Lines are drawn to guide the eye. P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 9 - P R O P E R T I E S O F P T F E P A S T E E X T R U D A T E S 9 4 In a die having a larger reduction ratio, resin particles are squeezed against one another to a greater extent over a longer flow path, resulting i n a higher extrusion pressure. This allows a greater extent o f mechanical interlocking to occur between adjacent particles. A greater amount o f fibrils is consequently created, as these interlocked crystallites are unwound while exiting the die. P rov ided that the quality o f the fibrils is unaffected by the increase i n the die reduction ratio, this translates into a mechanically stronger extrudate. However , this is only partially observed i n F ig . 9.4. B e y o n d a certain reduction ratio, the extrudate tensile strength is approximately unchanged or, as i n the case o f resin 2, decreases wi th further increase i n the die reduction ratio. T h e higher extrusion pressure associated wi th a die o f larger reduction ratio (see F ig . 9.4) causes the extrudate to be accelerated (spurted) out o f the die at a greater velocity. This results i n the breakage o f some fibrils i n the extrudate (as discussed above) and an overall reduction i n the fibril quality. The consequence is a mechanical weakening o f the extrudate. These competing effects between fibril quantity and quality affect the extrudate tensile strength i n opposing directions and become more important at larger reduction ratios, accounting for the observed trend i n F ig . 9.4. It is noted that the m a x i m u m die reduction ratio that is suitable for a particular resin is partly dependent on the molecular properties o f the resin. F o r example, as can be seen from Fig . 9.4, the extrudate obtained using resin 2 wi th a die o f reduction ratio o f 352:1 has a lower tensile strength than that obtained wi th a die o f reduction ratio o f 156:1. However , this is not observed wi th resin 4. It is also worthwhile to note again here that wi th resin 5, it is not possible to extrude the paste through a die o f reduction ratio o f 352:1. A s discussed in Chapter 8, such an extrusion attempt resulted i n a high velocity spurting o f periodically broken extrudates, due to the excessively high extrusion pressure. 9.3.2 Effect of Die Entrance Angle Figure 9.5 depicts the effect o f die entrance angle at constant throughput on the extrudate tensile strength for resin 3. T h e steady-state extrusion pressure corresponding to each experimental run is also shown i n the figure for comparison. T h e extrusion pressure generally increases wi th increasing die entrance angle, although at very small entrance angles P A S T E E X T R U S I O N O F P O L Y T E T O A I T X I O R O E T I - I Y L E N E F I N E P O W D E R RESINS CHAPTER 9 - PROPERTIES OF PTFE PASTE EXTRUDATES 95 (less than a ~ 7.5°), the frictional force becomes important and this trend reverses (see F ig . 8.9). A s discussed above, this increase i n the extrusion pressure translates into extrudates that are more fibrillated. However , while this is true, it appears that the overall quality o f the fibrils is increasingly compromised wi th the increase i n the die entrance angle. Th is is indicated by the decrease i n the extrudate tensile strength shown i n F i g . 9.5, contrary to an expected increase. as a. O ) c cu CO c CD CD CO " O =1 x UJ Resin 3 + 18wt.% ISOPAR G, 35°C R = 352:1, L/D = 0, Q = 75.4 mm3/s \ 36.2 MPa ^30.6 MPa 40.6 MPa : \ \ \ 43.4 MPa 56.4 MPa-10 20 30 40 Die Entrance Angle (a) 50 F i g . 9.5 The effect of die entrance angle on extrudate tensile strength. Also shown are the steady state extrusion pressures corresponding to each experimental run. Line is drawn to guide the eye. T h e increase i n the steady-state extrusion pressure wi th increasing die entrance angle is due to the paste flow path that becomes correspondingly less streamlined. Through visualization experiments, it has been shown previously that the paste flow pattern i n the die contraction zone becomes more "deformed" as the die entrance angle is increased. A s a consequence, the resulting extrudate exhibits a greater extent o f fibrillation. However , the degree o f fibril orientation and possibly, continuity, decreases. A greater number o f fibrils exiting the die w i l l be oriented i n directions other than the flow directions, due to the lesser streamlining o f the flow. This can be expected to result i n a larger swelling o f the extrudate. Indeed, diameter measurements o f extrudates obtained using dies o f various entrance angles PASTE EXTRUSION OF POLYTElTcAFLUOROElWLENE FINE POWDER RESINS C H A P T E R 9 - P R O P E R T I E S O F P T F E P A S T E E X T R U D A T E S 96 indicate this effect and this is shown i n F ig . 9.6. It can be seen from F i g . 9.6 that extrudate obtained using a die o f smaller entrance angle exhibits less swell, due to the greater streamlining o f the flow. O n the other hand, extrudate obtained using a die having a larger entrance angle exhibits greater swell. The overall effect is a reduction i n the fibril quality as the die entrance angle is increased, and consequendy a mechanically weaker extrudate is obtained (Fig. 9.5). 1.1 I i i i i | i i i i | i i . i | i i i i | i i i i I 1.0 | 0.9 <g <5 e .2 0.8 Q £ to "O I 0.7 h UJ 0.6 0.5 / O / Resin 3 + 18wt.% ISOPAR G, 35°C R = 352:1, L/D = 0, Q = 75.4 mm3/s 10 20 30 40 Die Entrance Angle (a) 50 Fig. 9.6 The effect of die entrance angle on extrudate diameter (Da = 0.508 mm). Line is drawn to guide the eye. 9.3.3 Effect of Die Aspect (L/DJ Ratio Fig . 9.7 shows the effects o f die L/Da ratio o n the tensile strength and diameter o f paste extrudates. It can be seen that increasing the die L/Da ratio increases the mechanical strength o f the extrudate. This is due to the improvement i n the overall fibril quality (orientation), as opposed to the creation o f more fibrils. W i t h the increase i n the die capillary length, there is sufficient time for the fibrils to relax and become more orderly i n their orientation. The creation o f fibrils, on the other hand, occurs only i n the die entrance zone (see F ig . 7.2), where the interlocking o f particles is possible. In the die capillary zone, the fibrillated paste flows i n a p lug f low manner, and the pressure drop across the capillary P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 9 - P R O P E R T I E S O F P T F E P A S T E E X T R U D A T E S 97 length is due mainly to the fact ional losses at the die wall . W i t h this consideration, one is, therefore, led to believe that extrudates produced using dies o f the same reduction ratio and entrance angle w i l l exhibit the same extent o f fibrillation, regardless o f the length o f the die capillary zone. 45-CO c OL 6 o> 5 CD 55 a> 4 'i/i c * 3 cu o IS TJ 3 * 2 LU 1 I I " f f l \ Resin 4 + 18wt.% ISOPARG, 35°C \ / R = 352:1, a = 45°, Q = 75.4 mm3/s \ i \ i \ / h \ / -•©-- Extrudate Tensile Strength -•A— Extrudate Diameter 1.1 - H o 0.9 | £ CD E 0.8 .2 a to •a 0.7 | LU 0.6 0.5 10 20 30 Die L/Da Ratio 40 50 F i g . 9.7 The effect of die L./Da ratio on extrudate tensile strength and diameter (Da = 0.508 mm). Lines are drawn to guide the eye. T h e role o f the die capillary zone i n improv ing the quality o f the fibrils i n the extrudate is somewhat evident f rom F ig . 9.7, where one can see that increasing the die L./Da ratio suppresses the extent o f extrudate swell. This implies that by increasing the capillary length, spurting o f fibrils near the exit o f the die entrance (contraction) zone is contained wi th in the capillary channel. Consequendy, the randomness o f fibril orientation is reduced and the overall fibril quality in the extrudate is improved. Visual ly, this is also apparent i n F ig . 9.8. Extrudates obtained using a die o f L/Da = 0 (Fig. 9.8 (a)) often show a fibrous appearance o n the surface, wh ich is visible even to the naked eye. This is due to the presence o f broken fibrils that are chaotically oriented, as they are spurted out o f the die at a high extrusion pressure. The extrudate appearance is vastly improved when a die having an L/Da > 0 is used (Fig. 9.8 (b)). This explains the increase i n the extrudate tensile strength as the die L/Da P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 9 - P R O P E R T I E S O F P T F E P A S T E E X T R U D A T E S 98 ratio is increased (Fig. 9.7). However, it is noted that since this effect occurs near the exit of the die contraction zone, beyond a certain capillary length, there will be no further improvement in the overall fibril quality, as shown by the plateau in the curves in Fig. 9.7. Apparently, using a die having L/Da ~ 20 will optimize the process in terms of product quality. (b) Fig. 9.8 Pictures of extrudates (resin 5) obtained using dies of (a) L/Da = 0 and (b) L/Da = 10 under the same experimental conditions. Note the visual difference in the extrudate surfaces. The same effect was observed with other extrudates. P A S T E E X T R U S I O N O F P O L V r E T R A F L U O R O E T I I Y L E N E F I N E P O W D E R RESINS C H A P T E R 9 - P R O P E R T I E S O F P T F E P A S T E E X T R U D A T E S 99 9.4 Effect of Resin Molecular Structure The effect o f resin melt creep viscosity (molecular weight) o n the extrudate tensile strength is shown i n F ig . 9.9. T h e results for bo th the homopolymer and copolymer series are plotted i n the figure. It can be seen that increasing the resin molecular weight increases the mechanical strength o f the extrudate for resins i n the same series. However , one has to be cautious before extending this conclusion across different polymer series. Further studies invo lv ing more samples are necessary before a general conclusion can be made. 14 CO CL 12 h 10 £ 8 h CO 0) co c CD UJ 4 h 18wt.% ISOPAR G,50°C • L/Da = 20, a = 45°, R = 56:1, Q = 75.4 mm3/s / A L/D„ = 20, a = 30°, / R = 336:1, Q = 124.9 mm3/s / A / 1 Melt Creep Viscosity x 10 (Pa.s) Fig. 9.9 The effect of resin melt creep viscosity (molecular weight) on extrudate tensile strength. Lines are drawn to guide the eye. Filled symbols represent copolymer series and unfilled symbols represent homopolymer series. It is worthwhile not ing that the steady-state extrusion pressure generally increases wi th increasing resin molecular weight. In some cases, it is not even possible to extrude a paste, due to its high molecular weight. F o r example, recalling the above discussion, an extrusion attempt using resin 5 at a reduction ratio o f 352:1 resulted i n the spurting o f broken extrudate pieces out o f the die i n a discontinuous fashion, due to the excessively high pressure that broke the fibrils. The increase i n the extrusion pressure wi th resin molecular weight, however, is caused by the 'hardness' o f the higher molecular weight resin particles P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 9 - P R O P E R T I E S O F P T F E P A S T E E X T R U D A T E S 100 (making them less deformable — see Chapter 6), rather than to the creation o f more fibrils. I n fact, the generally larger particle size o f higher molecular weight resin implies that there w i l l be a reduction i n the contact area between adjacent particles and consequendy, it is more reasonable to expect a smaller extent o f fibrillation. T h e fact that these particles are also less deformable further contributes to this. Furthermore, a higher molecular weight resin particle is made up o f a lesser number o f molecules, although they are longer (bulkier). Hence , there w i l l be fewer molecules to be unwound to create fibrils. Prel iminary analysis using D S C has also indicated this. The difference i n the first heat o f melt ing for a higher molecular weight paste tends to be smaller than that for a lower molecular weight paste. Th is can be seen i n F ig . 9.10, and is also apparent i n F i g . 9.1. This implies that the state o f crystallinity i n the higher molecular weight case is only slightly disturbed (i.e. less fibrillation). 0.07 0.06 l 0 ' 0 5 £ •o S 0.04 E I < f 0.03 E I <. 0.02 0.01 0 1 2 3 4 Melt Creep Viscosity x 10 ' 1 0 (Pa.s) Fig. 9.10 The difference in the degree of crystallinity between unprocessed and processes pastes having different molecular weights (melt creep viscosity). Recall that a high melt creep viscosity implies a high resin molecular weight. Extrudates obtained from a lower molecular weight resin should, therefore, exhibit a better mechanical (tensile) property, since they are more readily fibrillated, as discussed above. However , this is not observed i n F ig . 9.9. It appears that, although extrudates obtained from a lower molecular weight paste are composed o f more fibrils, these fibrils are i i i i | i i i i | 1 1 1 1 1 1 1 1 r - * - R = 156:1 j i i i I i i i i I i i i i L P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 9 - P R O P E R T I E S O F F I F E P A S T E E X T R U D A T E S 101 weaker compared to those obtained using a higher molecular weight paste. T h e difference i n the strength o f these fibrils may be due to the fact that, i n the higher molecular weight case, molecules are longer. T h e fibrils created are, therefore, more continuous, as compared to the shorter fibrils associated wi th the lower molecular weight paste. These fibrils remain intact during the flow and this evidently results i n a greater extrudate tensile strength, as can be seen i n F ig . 9.9. Hence, although a lower molecular weight paste extrudate is more fibrillated, the fibrils formed are relatively easy to break. 9.5 Effect of Lubricant Concentration T h e effect o f lubricant concentration o n the extrudate tensile strength is depicted i n F ig . 9.11 for resin 3. It can be seen that there is an op t imum lubricant concentration at wh ich the extrudate tensile strength is maximized. A n extrusion wi th a lubricant level less than the op t imum results i n an extrusion pressure that is too high which breaks the fibrils, although more fibrils are formed due to the increased friction between the resin particles. However , when the lubricant concentration is unnecessarily high, the extrudate becomes excessively wet and weak, sometimes not being able to ho ld its shape, due to the excessive local slippage between flowing particles in the die that results in a lesser number o f fibrils being formed. 3 CO o_ £ 2 D ) a CD i— 55 '55 c CD I— CD 1 1 3 i— X LU 0 15 16 17 18 19 20 21 22 23 Lubricant Concentration (wt.%) Fig. 9.11 The effect of lubricant concentration on extrudate tensile strength. Line is drawn to guide the eye. P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y U I N E F I N E P O W D E R RESINS C H A P T E R 9 - P R O P E R T I E S O F P T F E P A S T E E X T R U D A T E S 102 A l t h o u g h the op t imum lubricant concentration is expected to depend o n the average particle size o f the resin (molecular structure), this is not expected to be a major practical issue, since the average particle sizes o f paste extrusion grade resins are typically very similar. 9.6 Conclusions In this chapter, the effects o f die design, resin molecular structure and lubricant concentration o n the mechanical properties o f paste extrudates, and their relationship to the quantity and quality o f the fibrils formed during the extrusion process are discussed. Quantif ication attempts o n the extent o f fibrillation i n an extrudate using D S C and Raman spectroscopy, i n order to illustrate the issue o f fibril quantity and quality, is also discussed. It was found that a balance between fibril quantity and quality is necessary to ensure industrially acceptable products. T h e extent o f fibrillation i n an extrudate can be increased by performing extrusion using a die o f larger reduction ratio or entrance angle, or by decreasing the lubricant content i n the paste matrix, wh ich essentially increases the extrusion pressure level. However , too high an extrusion pressure tends to break the fibrils and is, therefore, detrimental to the quality o f the extrudate. B y increasing the die entrance angle, the quality o f the fibrils formed is also somewhat compromised, due to less streamlining o f the flow. Increasing the length o f the die capillary zone (or L/Da ratio) was found to increase the extrudate tensile strength, due to the increased degree o f fibril orientation i n the extrudate. T h e extrudate was also found to be visually more attractive (less fibrous) and exhibit less swelling. It was also noted that the pressure drop across the capillary length o f the die does not contribute to the formation o f more fibrils. Finally, it was found that a higher molecular weight resin produces stronger fibrils that account for the mechanical superiority o f the extrudate. P A S T E E X T R U S I O N O F P O L Y T E T R J V F L U O R O E T H Y L E N E F I N E P O W D E R R E S I N S CHAPTER 10: CONCLUSIONS 1 0 . 1 PTFE Paste Extrusion: Project Wrap - Up In this work, fundamental research has been performed o n paste extrusion as a unique processing technique for P T F E resins. P T F E paste extrusion takes advantage o f the near ambient transition temperatures o f P T F E , wh ich allows the resins to be fibrillated when sheared. The paste is essentially a mixture o f P T F E fine powder resin and a lubricating l iquid , wi th a typical l iquid concentration o f 16 wt .% to 25 wt.%. D u r i n g the process, the paste is initially preformed and later, extruded through a conical entry die using a ram extruder. D u e to the presence o f both solid and l iquid phases i n the paste matrix, the mechanism o f P T F E paste flow is fundamentally different f rom polymer melt flow. A typical extrusion pressure response curve for P T F E paste has three characteristic regions. T h e first region was found to be a manifestation o f the elastic nature o f P T F E paste, as it wets and fills the die conical zone. This is characterized by a peak i n the transient pressure curve. In the second region, the paste flow is at its steady state, and the extrusion pressure remains at an approximately steady level. I n the third region, the extrusion pressure gradually increases, as the paste i n the barrel becomes increasingly drier wi th time. This variation in the extrusion pressure has been found to result i n an extrudate having inconsistent properties. P T F E paste extrusion is carried out as a batch process. A n attempt to make the process continuous was found to be practically unattractive, since preforms produced separately do not continuously m o l d to one another i n the barrel dur ing extrusion. It was also found that, due to the nature o f the paste, analysis o f experimental data obtained using a modif ied extrusion procedure has to be performed wi th caution. F o r example, the extrusion pressure response at a particular extrusion rate may not represent the actual response at that condi t ion if, pr ior to this, the same paste has been extruded at a different piston speeds. This is due to the irreversible nature o f the paste properties, as caused by the high mobil i ty o f the l iquid phase (i.e. the lubricant) dur ing the extrusion process (liquid migration). 103 C H A F F E R 10 - C O N C L U S I O N S 104 In the preforming stage, P T F E paste is compacted to form a dense cylindrical billet. It was found that the preforming pressure has to be sufficiently high to ensure uni form compact ion along the preform length. Th is is especially true for a paste composed o f a higher molecular weight resin, wh ich was found to be physically "harder". D u r i n g the preforming o f such paste, pressure was dissipated close to the end at wh ich it was applied. Consequendy, the opposite end was less compacted and an insufficient dr iv ing force was present for the lubricant to migrate outwards. T h i s resulted i n a higher extrusion pressure required for this end o f the preform. Uni formi ty i n preform density can be improved by applying higher pressure, increasing the lubricant concentration i n the paste, or by compacting the paste over a longer period o f time. However , the latter two procedures were found to result i n the lubricant redistributing itself non-uniformly along the preform, wh ich is clearly undesirable. Th is leaves preforming pressure as the only adjustable parameter. T h e m i n i m u m pressure required to achieve a certain length o f uniformly compacted preform was found to correlate wel l wi th the resin S S G , wh ich is inversely related to the resin molecular weight. In this study, it was also shown that better uniformity can be achieved by applying pressure at both ends o f the preform. D u r i n g P T F E paste extrusion, fibrils are created. The presence o f fibrils i n the extrudates is clearly observable from S E M micrographs. A mechanism for fibrillation has been proposed, i n wh ich fibrils are considered as mechanically locked crystallites that are entangled in the die conical zone and unwound upon exiting the die. The consequence o f fibrillation is the formation o f a mechanically strong extrudate, wi th a degree o f crystaUinity lower than that o f the unprocessed resin. Th is was confirmed using D S C . B y performing visualization experiments, it was found that, i n the die conical zone where fibrillation occurs, the pattern o f paste flow can be described adequately by the "radial f low" hypothesis. T h e hypothesis assumes that all paste particles at the same radial distance from the virtual die apex are moving towards the die apex at the same velocity. Based on the radial flow hypothesis, a one-dimensional mathematical mode l was developed to describe the flow o f P T F E paste during extrusion. T h e model considers P T F E paste as an elasto-viscoplastic material that exhibits strain hardening during flow. A modif ied Kelv in ' s constitutive relation was used and a factional analysis was incorporated into the model derivation. T h e mode l is able to predict the effects o f extrusion rate and die design P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T I I Y L E N E F I N E P O W D E R RESINS C H A P T E R 10 - C O N C L U S I O N S 105 (die reduction ratio, L/Da ratio, and entrance angle) o n the steady-state extrusion pressure adequately, considering its simple form. The steady-state extrusion pressure was found to increase wi th increasing extrusion rate, and the die reduction and L/Da ratio. Increasing the die entrance angle initially decreases the steady-state extrusion pressure. Beyond a certain small entrance angle, however, the effect is reversed. It was also found that a higher molecular weight resin extrudes at a correspondingly higher steady-state pressure. I n order to produce commercial ly acceptable extrudates, adequate amounts o f fibrils w i th reasonable quality are required. F ib r i l quality can be described i n terms o f their continuity and degree o f orientation i n the extrudate. Raman spectroscopy was found to be useful i n an attempt to quantify fibril orientation. T h e quantity o f the fibril formed during the extrusion process, o n the other hand, can be inferred directly from the steady-state extrusion pressure. Overa l l , the extrudate tensile strength can be used to determine the state o f fibrillation i n an extrudate, since this variable takes into account the combined effects o f both fibril quantity and quality. I n this study, it was found that die design is critical i n affecting the extrudate quality. Increasing the die reduction ratio and entrance angle was found to increase the amount o f fibrils formed during the extrusion process, although the fibril quality was somewhat compromised. Increasing the length o f the die capillary zone was found to improve the extrudate appearance and mechanical strength. Th is is due to the increased degree o f fibril orientation i n the extrudate, rather than to the creation o f more fibrils. The extrudate was also found to exhibit less swell. Final ly, it was found that a higher molecular weight resin produces stronger fibrils that account for the mechanical superiority o f the extrudate. 10.2 Contributions to Knowledge Several novel contributions to knowledge have resulted f rom this research work. These are identified as follows. 1. The current commercial procedure for P T F E paste extrusion has been analyzed i n detail, and the physical significance o f each experimental aspect o f the process has been investigated. In addition, possible modifications to the process have been explored, i n an attempt to improve process efficiency. P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T I 4 Y L E N E F I N E P O W D E R RESINS C H A P T E R 10 - C O N C L U S I O N S 106 2. T h e preforming behavior o f P T F E pastes has been studied. T h e results provide an understanding o n h o w various operating variables affect the quality o f P T F E preform. A n empirical relationship that can be used to predict the m i n i m u m pressure required to produce a preform o f uni form density has been established. A n improvement i n the preforming procedure, wh ich involves the application o f pressure at both ends o f the preforming unit, has also been proposed and experimentally tested. 3. The mechanism o f flow involved i n P T F E paste extrusion has been investigated i n detail using S E M . A mechanism for fibrillation has been proposed and verified using D S C . Methods for quantifying fibrils using macroscopic extrusion pressure data, tensile strength, D S C and Raman spectroscopy have been illustrated. These are novel techniques that can be used for analysis o f other similar systems or processes. 4. The velocity pattern o f P T F E paste during extrusion has been studied. The "radial flow" hypothesis has been used to describe the flow pattern i n the die conical zone. Based o n this hypothesis, a one-dimensional mathematical model has been developed and used successfully to describe the rheology o f P T F E pastes. It may be possible to extend the use o f the mathematical model derived here to describe the flow o f other similar systems, such as highly filled polymeric systems. 5. T h e rheological behavior o f P T F E pastes has been studied experimentally. Effects o f various operating variables have been determined and discussed. T h e results can be used along wi th the developed mathematical mode l to improve die design and optimize the extrusion process, i n accordance wi th the rheological behavior o f an available resin. 6. The effects o f various operating variables o n the quality o f P T F E paste extrudates have been analyzed. The results provide an understanding o f the role o f fibrillation i n defining the final product properties. W i t h such understanding, it is possible to optimize the extrusion operating variables, i n order to produce extrudates that are commercially acceptable, wi th the ultimate objective o f reducing the amount o f process rejects. The results can also be used as a basis for resin selection, for a desired particular end use. M o s t importantly, this work has contributed to the fundamental knowledge o f P T F E paste extrusion, which is still at its infancy as far as research is concerned. Undoubtedly, P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 10 - C O N C L U S I O N S 107 more in-depth studies need to be performed i n the future to completely unravel the science behind the process. However , many o f the findings i n this work have provided significant initial steps towards a better macroscopic and microscopic understanding o f the process and, therefore, al lowed the commercial implementation o f P T F E paste extrusion to be carried out wi th greater confidence. Finally, many o f the analytical techniques employed here are novel , and can be used i n other studies involv ing processes similar to P T F E paste extrusion. 10.3 Recommendations for Future Work Several important aspects o f P T F E paste extrusion are yet to be studied. These are recommended below, as possible objectives for future research work. 1. T h e one-dimensional mathematical model developed i n this work to predict the f low o f P T F E pastes is dependent on several material constants. A l t h o u g h the effects o f resin molecular weight (melt creep viscosity) o n the values o f these constants have been assessed qualitatively, mathematical relationships that al low the predict ion o f the constant values, knowing a particular resin molecular structure, have yet to be established, primarily due to the l imited number o f samples available for this work. A more systematic study that involves a greater number o f resin samples wi th tailored molecular structures is recommended to establish such relationships. T h e final set o f mathematical relations wi l l then allow the determination o f a particular P T F E paste f low behavior, possibly without performing any experimental work. 2. A s has been discussed i n this work, lubricant migration is a significant phenomenon i n a process such as P T F E paste extrusion. A detailed study o n the subject is, therefore, warranted. T h e mechanism o f lubricant migration during P T F E paste extrusion needs to be determined. The effects o f various operating variables, such as die design, barrel diameter, resin particle size and distribution, and the lubricant concentration i n the paste mixture, should also be investigated. T h e results can be incorporated into the mathematical model to improve the predictions o f P T F E paste flow. 3. A detailed study o n the effects o f lubricant type, i n terms o f surface energy, viscosity, and volatility, o n the rheology o f P T F E pastes and the extent o f lubricant migration during P A S T E E X T R U S I O N O F P O L Y T E T O A F I J J O R O E T H Y L E N E F I N E P O W D E R RESINS C H A P T E R 10 - C O N C L U S I O N S 108 the extrusion process, w i l l provide a better insight into the role o f lubricant i n P T F E paste extrusion and is recommended for future work. 4. I n order to make the rheological study o f P T F E pastes more complete, the effects o f other variables, such as resin particle size and its distribution (if it is indeed possible to vary i n practice) and the surface roughness o f the die wall , should be investigated i n the future. T h e f low o f P T F E pastes through a hyperbolic die and a crosshead die for a wire coating process are also interesting and commercially useful subjects for future investigations. 5. In the fabrication o f sintered products f rom P T F E fine powder resins, the removal o f lubricant pr ior to sintering is an extremely crucial step. Therefore, it is desirable to study the drying characteristics o f extrudates, as functions o f their diameter and the type o f lubricant used i n the extrusion process. It is also important to study the temperature and density distribution i n the extrudate during sintering. Ultimately, a mathematical mode l may be developed to describe the sintering behavior o f P T F E paste extrudates. P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS NOMENCLATURE a, b, c B C E f L /* m M„ n N„N1a a R r, 0,0 r', 6', z c A H ^ A E U empirical parameters used to predict the effective length o f a preform, defined a s / . sin a/[2 . (1 - cos a)]. proportionality constant for the elastic term o f the mathematical model developed i n this work. A l s o refers to the proportionality constant i n Ludwik ' s power law model . die entrance and exit diameter, respectively. Db also refers to the barrel diameter. Young ' s modulus o f elasticity. Coulomb's law coefficient o f friction, between P T F E paste and die wall , length o f die capillary zone. effective length o f a preform, defined as the length o f a preform por t ion for wh ich compact ion density is approximately uniform. power law index for the viscous term o f the mathematical mode l developed i n this work. number average molecular weight. power law index for the elastic term o f the mathematical mode l developed i n this work. A l s o refers to the generalized power law index. first normal stress difference. Nla is the first normal stress difference calculated at the exit o f the die. volumetric flow rate. die reduction (contraction) ratio, defined as the ratio o f the die entrance to exit cross sectional area. average roughness spherical coordinate axes used i n force analysis i n the die conical (entrance) zone. I f used as subscripts, these indicate the directions i n wh ich a particular vector is acting. cylindrical coordinate axes used i n force analysis i n the die capillary zone. I f used as subscripts, these indicate the directions i n wh ich a particular vector is acting. radial distances as measured from the virtual die apex to the entrance and exit o f the die conical zone, respectively. time. constant o f integration. first and second heat o f melt ing (crystallization), respectively, as obtained from D S C . 109 N O M E N C L A T U R E 110 (X : die entrance angle. CC = 90" refers to flat die. £ p £ n , £ m : strains i n the three principal directions £, £ : strain and strain rate, respectively. Ymaxi 7ma* : m a x i m u m strain and strain rate, respectively. M a x i m u m strain is defined as the difference o f strains i n the first two principal directions. 7] : proportionality constant for the viscous term o f the mathematical mode l developed i n this work. A l s o refers to the viscosity coefficient i n Newtonian 's law for viscous flow. o" : stress. 0~„ 0 " n , o"HI : stresses i n the three principal directions, cr : yield stress. PASTE E X T R U S I O N OF P O L Y T E T R A F L U O R O E ' l T r Y L E N E FINE P O W D E R RESINS BIBLIOGRAPHY Adams , M . J . , I. A y d i n , B . J . Briscoe, and S. K . Sinha, A Finite Element Analysis of the Squeeze Flow of an Elasto-Viscoplastic Paste Material,). N o n - N e w t . F l u i d M e c h . , 71, 41-57 (1997). Amarasinghe, A . A . U . S., and D . I. W i l s o n , Interpretation of Paste Extrusion Data, Trans. Inst. C h e m . Eng . , 76A, 3-8 (1998). Amarasinghe, A . A . U . S., and D . I. W i l s o n , Statistical Analysis of Pressure Fluctuations in Paste Extrusion, The 1997 Jubilee Research Event : A T w o - D a y Symposium H e l d at the East Midlands Conference Centre, Not t ingham, 8-9 A p r i l 1997. Andr ikopou los , K . S . , D . Vlassopoulos , G . A . Voyiatz is , Y . D . Yiannapoulos and E . I . Kamitsos , Molecular Orientation of Hairy-Rod Polyester: Effects of Side Chain Length, Macromolecules , 31, 5465-5473 (1998). Ar iawan , A . B . , S. Ebnesajjad and S .G . Hatzikir iakos, Paste Extrusion of Polytetrafluoroethylene (PTFE) Fine Powder Resins, Rheo. A c t a , manuscript submitted for publication (2001). Ar iawan , A . B . , S. Ebnesajjad and S . G . Hatzikir iakos, Preforming Behavior of PTFE Pastes, Powder Tech. , manuscript accepted for publication (2001). Ar iawan , A . B . , S. Ebnesajjad and S . G . Hatzikir iakos, Properties of PTFE Paste Extrudates, P o l y m . E n g . Sci. , manuscript submitted for publication (2001). A S T M D l 710-96, Standard Spedfication for Extruded and Compression Molded Polytetrafluoroethylene (PTFE) Rod and Heavy-Walled Tubing (1996). A S T M D4895 , Standard Spedfication for Polytetrafluoroethylene (PTFE) Resin Produced from Dispersion (1997). Bai rd , D . G . , and D . I. Coll ias , Polymer Processing, But terworth-Heinemann, M A , 1995. Bates, A . J . D . , and J . Bridgwater, Modelling of Non-Uniform Paste Flow, T h e 1994 I C h e m E Research Event: A T w o - D a y Symposium H e l d at Universi ty College L o n d o n , Jan 5-6, 1994. Bates, A . J . D . , and J . Bridgwater, The Flow of Paste in Complex Geometries, T h e 1993 I C h e m E Research Event , Jan 6-7, 1993. Benbow, J . J . , and J . Bridgwater, Paste Flow and Extrusion, O x f o r d Universi ty Press, O x f o r d , 1993. Benbow, J . J . , and J . Bridgwater, The Influence of Formulation on Extrudate Structure and Strength, C h e m . E n g . Sci. , 42, 735-766 (1987). Benbow, J . J . , E . W . Oxley , and J . Bridgwater, The Extrusion Mechanics of Pastes — The Influence of Paste Formulation on Extrusion Parameters, Chem. E n g . Sci. , 42, 2151-2162 (1987). 111 B I B L I O G R A P H Y 112 Benbow, J . J . , N . Ouchiyama, and J . Bridgwater, On the Prediction of Extrudate Pore Structure from Particle Si%e, C h e m . E n g . C o m m . , 62, 203-220 (1987). Benbow, J . J . , T . A . Lawson , E . W . Oxley , and J . Bridgwater, Prediction of Paste Extrusion Pressure, Ceramic Bulletin, 68, 1821-1824 (1989). B ingham, E . C , Fluidity and Plasticity, M c G r a w - H i l l , N e w Y o r k , 1922. Blanchet, T . A . , Polytetrafluoroethylene, i n H a n d b o o k o f Thermoplastics, (O. Olabis i , ed.), 1 s t ed., Marce l Dekker , Inc., N e w Y o r k , 1997, 981-1000. Bouvard , D . , Densification Behaviour of Mixtures of Hard and Soft Powders Under Pressure, Powder Tech. , 111,231-239 (2000). B o w e r D . I . , and W . F . Maddams, The Vibrational Spectroscopy of Polymers, Cambridge Universi ty Press, N e w Y o r k , 1989. Bridgwater, J . , Paste Extrusion — an Overview, in Recent Advances in Chemical Engineering, Tata M c G r a w - H i l l , N e w D e l h i , 1989. Burbidge, A . S., J . Bridgwater, and Z . Saracevic, Liquid Migration in Paste Extrusion, C h e m . E n g . Res. Des ign , 73, 810-816 (1995). Burbidge, A . S., J . Feaey, A . W i l s o n , and J . Bridgwater, Liquid Phase Migration in Extrusion Pastes, Proceed, o f T h e W o r l d Congress o n Particle Technology, Br ighton , U K , July 6-9, 1998. Carley, J . F . , and E . v o n H o l t z , Flow of RX-08-FK High Energy Paste in Capillary Rheometer, J . Rheol . , 41, 473-489 (1997). Cavanaugh, R. J . , Private Communica t ion , August 1999. Chakrabarty, J . , Theory of Plasticity, M c G r a w - H i l l B o o k C o . , Singapore, 1998. Chevalier, L . , E . H a m m o n d , and A . Po i tou , Extrusion of Ti02 Ceramic Powder Paste, J . Mat . Process. Tech. , 71, 243-248 (1997). Clark, E . S., M U U S , L . T . Z . , Krist, 117-119, 1962. Cottrel l , T . L . , The strength of Chemical Bonds, 2 n d ed., Butterworths, Washington, D . C , 1958. Craig, R . F . , Soil mechanics, E & F N Spon, N Y , 1997. Dav i s , E . A . , and J . D u k o s , Theory of Wire Drawing,]. A p p l . Mech . , 11,193-198 (1944). Dealy, J . M . , and K . F . Wissbrun, Melt Rheology and its Role in Plastics Processing - Theory and Applications, V a n Nos t rand Re inho ld , N e w Y o r k , 1990. P A S T E E X T R U S I O N O F P O L Y l T i T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS B I B L I O G R A P H Y 113 Deimede V . , K . S . Andr ikopou los , G . A . Voyiatz is , F . Kons tandakopoulou and J . K . Kall i ts is , Molecular Orientation of Blue Luminescent Rigid-Flexible Polymers, Macromolecules , 32, 8848-8856 (1999). D o b a n , R. C , A . C . Knigh t , J . H . Peterson and C . A . Sperati, Molecular Weight of Polytetrafluoroethylene, D u P o n t C o . , presented at 130 t h Meet ing, A m e r i c a n Chemica l Society, Adan t ic Ci ty , N . J . , September 16-21,1956. D o m a n t i , A . T . J . , and J . Bridgwater, Extrusion Induced Defects, T h e 1996 I C h e m E Research Event : T h e Second Equropean Conference for Y o u n g Researchers i n Chemica l Engineering, A p r i l 2-3, 1996. D o m a n t i , A . T . J . , Surface Fracture in Paste Extrusion, P h . D . Dissertation, Department o f Chemica l Engineering, University o f Cambridge, Cambridge, 1998. Doraiswamy, D . , A . N . Munumdar , I. Tsao, A . N . Beris, S. C . Danfor th , and A . B . Metzner , The Cox-Mer\ Rule Extended: A Rheological Model for Concentrated Suspensions and Other Materials with a Yield Stress,}. Rheol . , 35, 647-685 (1991). D u P o n t Fluoroproducts , Molecular Weight of PTFE, Technical Bul le t in , Wi lming ton , Delaware, 2001. D u P o n t Fluoroproducts , Teflon® Fine Powder Processing Guide, Technical Bul le t in , Wi lming ton , Delaware, 1991. D u P o n t Fluoroproducts , Teflon® PTFE Fluoropolymer Resin — Processing G u i d e for Fine Powder Resins, Technical Bullet in , Wi lming ton , Delaware, 1994. D u P o n t Fluoroproducts , Teflon® PTFE Properties Handbook, Technical Bul le t in , Wi lming ton , Delaware, 1996. Ebnesajjad S., Private Communica t ion , October 2001. Ebnesajjad, S., Fluoroplastics, Vol. 1: Non — Melt Processible Fluoroplastics, The Definitive User's Guide and Databank, Plastics Des ign Library, W i l l i a m Andrew, Inc., 2000. E x x o n Corp . , Isopar® Solvents, Publicat ion D G - 1 P , 1994. Fayed, M . E . , and L . Ot ten (eds.), Handbook of Powder Science and Technology, Chapman and H a l l , N e w Y o r k , 1997. Feir ing, A . E . , J . F . Imbalzano, and D . L . K e r b o w , Advances in Fluoroplastics, Trends i n P o l y m . Sci. , 2, 26-30 (1994). Folda , T . , H . Hof fmann , H . Chanzy, and P . Smith, Liquid Crystalline Suspensions of Poly(tetrafluoroethylene) Whiskers', Nature, 333, 55-56 (1988). Fr i th , W . J . , J . Mewis , and T . A . Strivens, Rheology of Concentrated Suspensions: Experimental Investigations, Powder Tech. , 51, 27-34 (1987). P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS BIBLIOGRAPHY 114 Gangal , S. V . , Polytetrafluoroethylene, Homopolymers of Tetrafluoroethylene, i n Encyclopedia o f Polymer Science and Engineering, 2 n d ed., J o h n Wi l ey & Sons, N e w Y o r k , 1989, 577-600. Gangal , S. V . , Polytetrafluoroethylene, i n Encyclopedia o f Chemica l Technology, 4 t h ed., J o h n Wi ley & Sons, N e w Y o r k , 1994, 621-644. G o t z , J . , D . Mul le r , H . Buggisch, and C . Tasche-Lara, NMR Flow Imaging of Pastes in Steady-State Flows, C h e m . E n g . P r o c , 33, 385-392 (1994). Hatzikir iakos, S .G . , and J . M . Dealy, The effect of interface conditions on wall slip and melt fracture of high density polyethylene, S P E A N T E C '91 Tech . Papers, 2311-2314 (1991). Herschel , W . H . , and R . Buckley, KolloidZ., 39, 192 (1926). Hof fman , O . , and G . Sachs, Introduction to the theory of Plasticity for Engineers, M c G r a w - H i l l Company, 1953. Ho lmes , D . A . , and E . W . Fasig, Jr., Plunkett, R. J . , U S Patent 3,819,594, assigned to D u P o n t de Nemours and Company, June 1974. H o r r o b i n , D . J . , and R. M . Nedderman, Die Entry Pressure Drops in Paste Extrusion, C h e m . E n g . Sci . , 53, 3215-3225 (1998). J i r i , H . , T . Jana, and O . Frantisck, The Forming of Ceramic Paste, Ceramic-Silikaty, 33, 1-10 (1989). K a l i k a , D . S . , and D e n n , M . M . , Wall slip and extrudate distortion in linear low-density polyethylene, J . Rheol . , 31,815-834 (1987). K i m , C . J . , K . B . K i m , I. H . K u k , G . W . H o n g , S. D . Park, S. W . Yang , and H . S. Shin, Fabrication and Properties ofYba2Cu}07_s-Ag Composite Superconducting Wires by Plastic Extrusion Technique,]. Mat . Sci . , 32, 5233-5242 (1997). K i m , C . U . , J . M . Lee, and S. K . Ihm, Emulsion Polymerisation of Tetrafluoroethylene: Effects of Reaction Conditions on Particle Formation,]. Fluorine Chemistry, 96, 11-21 (1999a). K i m , C . U . , J . M . Lee , and S. K . Ihm, Emulsion Polymerisation of Tetrafluoroethylene: Effects of Reaction Conditions on the Polymerisation Rate and Polymer Molecular Weight, J . A p p l . P o l y m . Sci. , 73, 777-793 (1999b). Kobayashi , S., and E . G . , Thomsen , Upper and Tower Bound Solutions to Axisymmetric Compression and Extrusion Problems, Int. J . M e c h . Sci . , 7, 127-143 (1965). K u d o , H . , Some Analytical and Experimental Studies of Axi-symmetric Cold Forging and Extrusion — I, Int. J . M e c h . Sci. , 2, 102-127 (1960). Larson, R. , Constitutive Equations for Polymer Melts and Solutions, Butterworths, Bos ton , 1988. PASTE EXTRUSION OF POLYFETRAFLUOROETHYLENE FINE POWDER RESINS B I B L I O G R A P H Y 115 Lehnert R. J . , P.J. Hendra , N . Evera l l and N . J . Clayden, Comparative Quantitative Study on the Crystallinity of Polytetrafluoroethylene) including Raman, Infra-Red and 19F Nuclear Magnetic Resonance Spectroscopy, Polymer , 38, 1521-1535 (1997). Lewis E . E . , and C M . Winchester, Rheology of Lubricated Tetrafluoroethylene Compositions, Ind. E n g . Chem. , 45,1123-1127 (1953). Lewis , E . E . , and C . M . Winchester, Rheology of Lubricated Polytetrafluoroethylene Compositions — Equipment and Operating Variables, Ind. E n g . Chem. , 45, 1123-1127 (1953). L i , Y . Y . , and J . Bridgwater, Prediction of Extrusion Pressure Using an Artificial Neural Network, Powder Tech. , 108, 65-73 (2000). L o n t z , J . F . , and W . B . Happo ld t Jr., Teflon® Tetrafluoroethylene Resin Dispersion, Ind. E n g . Chem. , 44, 1800-1804 (1952). L o n t z , J . , J . A . Jaffe, L . E . R o b b and W . B . Happoldt , Jr., Extrusion Properties of Lubricated Resin from Coagulated Dispersion, Ind. E n g . Chem. , 44, 1805-1810 (1952). Louge, A . , A Two-Phase Model for Paste Plow Including Filtration and Segregation Effects, Comptes Rendus de LaAcademie des Sciences, Serie II, 322, 785-791 (1996). Ludwik , P. , Elemente der Technologischen Mechanik, Springer-Verlag, Ber l in , 1909. Mac leod H . M . , and K . Marshal l , The Determination of Density Distributions in Ceramic Compacts Using Autoradiography, Powder Tech . 16, 107-122 (1977). Malamataris, S., and J . E . Rees, Viscoelastic Properties of Some Pharmaceutical Powders Compared Using Creep Compliance, Extended Heckel Analysis and Tablet Strength Measurements, Int. J . Pharm. , 92, 123-153 (1993). Mazur , S., Paste Extrusion of Polytetrafluoroethylene) Fine Powders, i n Polymer Powder Technology ( M . Nark is and N . Rosenzweig, eds.), J o h n Wi ley and Sons, N e w Y o r k , 1995, 441-481. M c A d a m s , J . , Private Communica t ion , A p r i l 2000. M c A d a m s , J . , Private Communica t ion , August 1999. Michae l i , W . , Extrusion Dies for Plastics and Rubber, Hanser, N e w Y o r k , 1992. O ld royd , J . G . , A Rotational Formulation of the Equation of Plastic Flow for a Bingham Solid, Proc . Cambridge Philos. S o c , 43, 100-105 (1947). Papanastasiou, T . C , Flow of Materials with Yield,}. R h e o l , 31, 385-404 (1987). Pashias, N . , D . V . Boger, A Fifty Cent Rheometerfor Yield Stress Measurement,J. Rheol . , 40, 1179-1189 (1996). P A S T E E X T R U S I O N O F P O L Y T E T R A F L U O R O E T H Y L E N E F I N E P O W D E R RESINS B I B L I O G R A P H Y 116 Plunkett, R. J., The History of Polytetrafluoroethylene: Discovery and Development, in High Performance Polymers: Their Origins and Development, Proceed, of Symp. on the Hist, of High Perf. Polymers at the ACS Meeting in New York, April 1986, (R. B. Seymour and G. S. Kirshenbaum, eds.), Elsevier, New York, 1987. Plunkett, R. J., US Patent 2,230,654, assigned to DuPont Co., February 4,1941. Popovich, L. L., D. L. Feke, and I. Manas-Zloczower, Influence of Physical and Interfacial Characteristics on the Wetting and Spreading of Fluids on Powders, Powder Tech., 104, 68-74 (1999). Ramamurthy, A.V., Wall slip in viscous fluids and influence of materials of construction, J. Rheol., 30, 337-357 (1986). Ring, T. A., Fundamentals of Ceramic Powder Processing and Synthesis, Academic Press, New York, 1996. Saint — Venant, B., Memoire sur I'etablissement des Equations Differentielles des Mouvements Inte'rieurs Opens dans les Corps Solides Ductiles, Compt. Rend. Acad. Sci. Paris 70, 473 (1870). Schultz, M . A., and L. J. Struble, Use of Oscillatory Shear to Study Flow Behavior of Fresh Cement Paste, Cement and Concrete Research, 23, 273-282 (1993). Sheppard, W. A., and C. M . Sharts, Organic Fluorine Chemistry, W. A., Benjamin, Inc., New York, 1969. Shimizu, M . , C. Ikeda, and M . Matsuo, Development of High Modulus and High Strength Polytetrafluoroethylene) Fibers by Elongation at Liquid Crystalline State, Macromolecules, 29, 6724-6729 (1996). Snelling G. R , and J. F. Lontz, Mechanism of Lubricant-Extrusion of Teflon® TFE-Tetrafluoroethylene Resins,]. Appl. Polym. Sci. I l l , 257-265 (1960). Sperati, C. A., Physical Constants of Fluoropolymers, in Polymer Handbook, (J. Brandrup and E.H. Immergut, eds.), John Wiley and Sons, New York, 1989. Spitteler, P. H . J., L. F. Gladden, A. M . Donald, and J. Bridgwater, ESEM and NMR Imaging of Water Based Ceramic Pastes, Proceed, of World Congress on Particle Technology, Brighton, U K , July 6-9, 1998. Suwa, T., M . Takehisa and S. Machi, Melting and Crystallisation Behavior of Poly (tetrafluoroethylene): New Method for Molecular Weight Measurement of Polytetrafluoroethylene) using a Differential Scanning Calorimetry,] of Appl. Polymer Sci., 17, 3253-3257 (1973). Tordella, J.P., Fracture in the extrusion of amorphous polymers through capillaries, J. Appl. Phys., 27, 454 (1956). Tuminello, W. H. , T. A. Treat, and A. D. English, Polytetrafluoroethylene): Molecular Weight Distributions and Chain Stiffness, Macromolecules, 21, 2606-2610 (1988). P A S T E E X T R U S I O N O F P O L Y T C T U A F L U O R O E T I - r Y L E N E F I N E P O W D E R RESINS B I B I J O G R A P H Y 117 Vandeneede, V . , G . Moortgat , F . Cambier , Characterisation of Alumina Pastes for Plastic Moulding,}. E u r o . Ceramic S o c , 12, 225-231 (1997). Voyiatz is , G . A . , G . Petekidis, D . Vlassopoulos, E . I . Kamitsos and A . Bruggeman, Molecular Orientation in Polyester Films Using Polarised Laser Raman and Fourier Transform Infrared Spectroscopies andX-Ray Diffraction, Macromolecules , 29, 2244-2252 (1996). Voyiatz is , G . A . , Private Communica t ion , M a y 2000. W i l d m a n , R . D . , S. Blackburn , and D . J . Parker, Investigation of Paste Flow Using Positron Emission Particle Tracking, Powder Tech. , 103, 220-229 (1999). Y u , A . B . , J . Bridgwater, A . S . Burbidge, and Z . Saracevic, Liquid Maldistribution in Particulate Paste Extrusion, Powder Tech. , 103,103-109 (1999). Yurchenko , E . N . , V . Y . Prokofev, A . P . I l lyin, and Y . G . Shyrokov, Optimum Properties of Extrudate Paste Powders on the Basis of Ti02> Al203, and Other Substances for Production of Honeycomb Supports, Reaction Kinet ics and Catalysis Letters, 60, 269-277 (1997). Zhang , D . Z . , and R. M . Rauenzahn, A Viscoelastic Model for Dense Granular Flows, J . Rheol . , 41,1275-1298 (1997). Z i sman , W . A . , Surface Properties of Plastics, Record o f Chemical Progress, 26 (1965). P A S T E E X T R U S I O N O F POLY1 J T F R A F L U O R O E T F I Y L E N E F I N E P O W D E R R E S I N S 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0058968/manifest

Comment

Related Items